 Good morning from Stanford University. I'm Will Chu. I'm the co-director of the Storage X Initiative and together with my colleague, Professor Itui, I'm very pleased to welcome everyone to the final symposium in the spring quarter for the Storage X symposium. We'll finish this outstanding quarter with a talk on cathode materials chemistry and electrochemistry. Cathode is a topic that's very near and close to my heart. I think it is a very rich playground for doing fundamental science and innovation and I'm extremely pleased to have an outstanding panel today with speakers and participants from three continents to finish off the quarter. Joining us from Germany is Professor Hubert Gasthiger from the Technical University of Munich and I'll introduce him in more detail in a bit. And then joining us very late in the evening from Korea is Professor Young-Kook Sun from Hongyang University. Both of them have done seminal and significant work to advance the cathodes making what they are today. And let me go ahead and start introducing our first speaker so we can get started. Hubert is, holds the chair of electrochemistry at the Technical University of Munich and he is truly a broadly trained individual. He's done it all. He started his career at Lawrence Berkeley National Labs just right here in the Bay Area after receiving his PhD at UC Berkeley. Then he spent a decade working in industry at General Motors developing electric catalysis and since 2009, 2010 he holds the chair of electrochemistry in Munich. I cannot think of a more diversely knowledgeable person than Hubert on anything electrochemistry and certainly his contribution extends beyond batteries but also fuel cells and electric catalysis. We are really pleased to have him today to talk about the interfacial electrochemistry of cathode materials. Hubert if I can have you come to the stage. Okay. There you go Hubert. We are very much looking forward to your lecture. So thank you very much for this opportunity to present a talk at this wonderful symposium and thank you very much Will and E for the invitation. I'm also glad I'm the first one because it would be rather intimidating to be after young cook so I'm glad I can start. So what I want to talk about the next half hour is about the different degradation phenomena in nickel rich cathode active materials so called NCMs and so this is a talk we put together with three students of mine and is a little bit of a review of work we have done in this area. So why a nickel rich NCMs of course at this symposium I think I do not need much of an introduction but essentially what happened in the last years was that the nickel content was gradually increased from one on one six to two eight one one and now even higher nickel content materials increasing the capacity and then in recent years also there has been attempts to gradually increase the upper cutoff voltage so it is of course interesting to see what the degradation phenomena of these different materials are at high nickel contents and at high cathode potentials. So the materials we talk about are these layer transition metal oxides with the very small excess of lithium and a theoretical capacity of 275 million miles per gram if you were to extract all of the lithium. So if we look at the energy density so the energy density increases with nickel content so if we look here at the potential of 4.3 volts versus lithium so this would be 4.4 volts in a full cell 4.2 volts in a full cell so you go from a NCM 111 to 811 it's a significant increase in capacitance and the same is true of course at 4.5 volts where you can get significant gains. So if we look at these different materials in terms of the cost there is a paper from the mentoring group which looked at a potential versus lithium at 4.2 volts and then showed that the cost of the nickel rich materials is lower of course predominantly because of the reduced amount of cobalt. However the drawbacks of these materials have to do with their higher sensitivity so when you increase the nickel content the stability towards the ambient towards moisture specifically decreases also the thermal stability decreases so this was actually shown by Professor Sun in a predeterminal paper and then also the cycling stability decreases and so what we want to look at in this talk is what are sort of the different degradation mechanisms of the cathode material so focusing really on the cathode material during battery operation and so if we look at the different aging mechanisms with respect to the cathode active materials there is a few things which are pretty well known and so when we look at an electrochemical cell what we have is that particularly at higher potentials the NCM materials release oxygen we have a surface reconstruction and we create an electrochemically inactive phase. In addition during this oxygen release we oxidize the electrolyte our hypothesis was that it goes via singlet oxygen but there's of course other hypotheses in the literature but nevertheless what happens there is that you produce acidic species typically HF which leads to transition metal dissolution and migration to the cathode. The other thing we have is depending on the cycling potential the time of cycling we see a cracking of polycrystalline NCM materials so what we get is a gradual break up of the secondary agglomerates into their primary crystallites and that in principle can lead to electronically isolated phases in the center of the secondary agglomerates and also of course at high potential we have the oxidation of the electrolyte which again produces protic species which get reduced on the negative electron leading to hydrogen and then of course we have a third mechanism which is the so-called lithium nickel mixing particularly well known for L and O and the hypothesis there is of course that these occupied lithium sites by the transition metal lead to slower diffusion. So the question is how do we do a study where we exclusively focus on the cathode active materials and so what we do is we try to use we use a pre-lithiated N-out so in order to eliminate constraints which come from the active lithium loss on the N-out electrode we use a non-gassing N-out in order to essentially eliminate the contributions from the N-out to the gas signals and we use an excess of electrolyte in order to not have to consider the continuous consumption of the electrolyte. If we look at this what can we measure when we look at the formation of a surface reconstructed phase we expect an increase in the charge transfer resistance which in principle one would be able to measure in the case of particle cracking what we expect is an increase of the electrochemically active surface area meaning of the electrolyte wetted surface area so we would increase expect an increase in the charge transfer resistance actually a little benefit of this is that the sorry in the capacitance a little benefit is that the charge transfer resistance also would go up a little bit and in case of lithium nickel mixing of course we expect that we have a slower slower lithium diffusion coefficient in the active material in the bulk of the material so how do we deconvolute the impedance signals of course what we can measure is the impedance of the full cell but if we want to focus on the cathode active material we can either do impedance with a micro reference electrode so then we can collect the or acquire the impedance spectra of the cathode active material by itself the other thing is we can do gcr pulses in cells when we have a lithium reference electrode and then essentially get the impedance of the cathode or we can do impedance spectroscopy which I also showed for a counter electrode which has a very very low impedance so that they can be ignored so these are the different approaches we use in our studies so the talk is split in three sections one section relates to oxygen release and surface reconstruction the other one has to do with particle cracking and the third one with potential this image of mixing so the oxygen release we typically measure with online electrochemical mass spectroscopy OEMS and the studies we have conducted in the past through this regard was a procedure where we increased the potential in the linear scan and we measured the you know cell the cell voltage the cathode potential is the function of the state of charge this is shown here for three different materials 111 a 622 and an 811 material and as a comparison a high voltage demand and so what we see is that for the ncm materials we have the evolution of oxygen at approximately 80% state of charge and that is actually irrespective of the nickel content when we look at potential then of course what you see is that the oxygen release on the different materials occurs later for the nickel pool materials then for the little rich material right so this has to do of course with the slightly lower voltage of the nickel rich materials and the lm o which here goes to as high potential as the ncm so pool here so about five volts we see absolutely no oxygen right and so what we know is that the spinel structure is actually stable against the oxygen release when we look at other gases what we see is that whenever we have the onset of oxygen evolution we also have the onset of co and co2 evolution which we ascribe to the degradation of the electrolyte and as I said in our studies we ascribe it to the reaction of the release the singlet oxygen with the electrolyte but as I said there's other hypothesis in the literature and of course upon oxygen release you get a material which is poorer in oxygen than the layered structure and what one finds for nickel rich materials is that it reconstructs into a rock salt like face and so this is some data from the lab from Jürgen Jannik and Bella of bsf for a nickel rich material so 851005 after 500 titles whether they really identify rock salt faces now when we look at the the material morphology what we know is that these polycrystalline ncm are actually composed of secondary agglomerates in the order of maybe 5 to 10 micrometers in diameter and the secondary agglomerates consist of primary crystallites which are in the order of 0.2 micrometers so if we measure the BET surface area of these materials we get about 0.3 meters per gram this is what you would measure and if you use a surgery approximation you would calculate an effective particle diameter of about four microns on the other hand if the material were to correct during cycling or during other treatments into the primary crystallites which are about 0.2 microns in diameter the spherical approximation would say that the BET surface area increases to about six meters per gram right and so the truth of course is somewhere in between you start out definitely with a low BET but what we find is that typically the BET area increases over the course of cycling and this is what we will look at in the next section which is about the particle cracking and in order to develop an in-situ method to follow the particle cracking we first of course had to do some sort of method method validation and the method validation we did was by just simply compressing cathodicative materials to crack it by mechanical force and so what we know from Fibre-CM measurements is that these NCM materials are pretty dense have a few occluded volumes but in general otherwise they are rather dense on the other hand if you compress the material at high compressive forces you can see the cracking of the materials of course these are very high forces here as I said this was really just a validation measurement so the question is can we quantify this because of course Fibre-CM is beautiful because you can see the images and it is very visual and very clear but they cannot be quantified so what we try to do here is to develop a method based on BET but using krypton instead of nitrogen for the BET measurements as krypton there's about two orders of magnitude more sensitive now when you measure powders that of course is irrelevant you could just use enough powder to conduct a nitrogen BET but when you use electrodes then of course your surface area your total mass is too little and so in our case it only works really when we use krypton BET so what we measure here is the krypton BET of the pristine powder when it is not compressed sorry of an electrode when it is not compressed and when we compress the electrode what we see is that the BET the krypton BET surface area increases now of course this is a total area of the electrodes so this contains also the conductive carbon which has a very high BET surface area itself compared to the cathodic materials so what we have to do is of course subtract the contribution from the conductive carbon and the electrode and so this was done by a model electrode which only contained the binder and the conductive carbon and so what we can see is that the remainder here is the BET surface area of the cam in the electrode and this is actually very well with the powder here and as we compress this we see that the BET surface area increases as we crack these primary secondary agglomerates more and more so what we find right of course nothing unexpected compression reduces cracks and this increases the electrochemical surface area and so we use this to actually calibrate our new method which was trying to extract the capacitance of the cathode electrode using a transition a transmission line model with constant phase elements and when we do this we utilize so-called blocking conditions and so blocking conditions can be obtained by either using a non-intercalating electrolyte which was done in this example here or by going into conditions where the charge transfer resistance becomes very very large and so the typical impedance spectrum of blocking conditions is essentially more or less a vertical line not quite vertical so the constant phase element exponent is 0.88 instead of 1.0 which would be for capacitor and then you see here a little bit of a contact resistance so when we do this for all of these electrodes we can acquire the impedance data and we can analyze the impedance data in terms of analyzer capacitance from the impedance data and then what we get here is exactly the dash bars here and what we can show is that there is actually a very good agreement between the capacitance increase and the increase in the BET surface area so we can use the measure the capacitance of the electrode of the cathode electrode as a measure of the effective BET area increase of course we have to do the same subtraction of the capacitance contributed by the carbon so then we use this method to try to follow the particle cracking upon cycling so the first example here is done by XZ2 utilizing a krypton BET measurement so what we know is when we cycle NCM particles we have a significant change in the lattice parameters and then the lattice volume and so one expects to have a cracking of these particles due to mechanical strain which occurs particularly at the interfaces between the primary crystallites of a polycrystalline material so when we cycle the material this goes to 4.2 volts what we can see after the first charge in the 5th SCM is that you can now see the development of cracks in the material these cracks actually close again after the first few cycles when you go into the discharge state and when we so this this can be shown here you can see the close a little bit again of course very qualitative in this case and when we cycle for many cycles then of course even in the discharge state you see the cracking of the material quite clearly so the cracking is actually irreversible. If we now measure the BET surface area of materials what we can do is we can measure the BET surface area of the pristine powder as I said before it's about 0.3 meter square per gram for these materials we can also measure the BET surface area of a charged electrode that of course can only be done with krypton BET because you can not have so much active mass in your instrument and then we can actually follow the BET surface area of cycled electrodes over the course of cycling and this of course is very tedious because doing 5th SCM is a big effort plus also cannot be quantified and the krypton BET of course requires a lot of experimentation to harvest electrodes put them in the BET and so on so what we wanted to do is to see whether we can follow this with our capacitance based method and so this is shown here on the next slide for an NCM material 6 to 2 NCM and we utilize for this a gold wire reference electrode to something we published before where essentially we insert the micro reference electrode between the two electrodes and we utilize a pre-lithiated LTO electrode because that allows us to bring the cathode electrode after each cycle into blocking conditions meaning completely charging the electrode where we know that the charge transfer resistance becomes very large so we get the almost semi-infinite charge transfer resistance and this allows us to get again a blocking condition signal from which it is very easy to quantify the capacitance so this is shown here for the first cycle so if we then cycle the material of course the number of cycles comes a little bit later these I believe 200 cycles what we can see is the change in capacitance this is just marked by the 180 millihertz frequency point and when we look over the cycling then what we see is that depending on the cutoff potential 3.9 volts versus lithium 4.1 or 4.2 or 4.5 we get a stronger increase in the capacitance so again as I said before we have to subtract the contribution from the pristine electrode and so the ratio of the pristine electrode minus the carbon contribution which is this part here and the signal afterwards that is our increase in active surface area which we deduce from these capacitance measurements and of course what we see is that higher potentials you get more cracking you expose more surface area you have more side reactions and this we believe is the reason why when you go to very high potentials you have more capacity fading for one of the reasons than if you cycle at lower potentials right so here the specific capacitance changes by maybe 10-20 percent and so this is of course well known and so the other effect we wanted to look at is what is the effect of oxygen released from the surface upon the surface area increase of the material and so this is shown here for NCM when we go to very high state of charge meaning beyond this 80 percent SOC where we know we get oxygen release so this one first one here is a polycrystalline material 622 and so the experiment we conduct here is that we gradually increase the potential in 100 millivolts steps we measure the capacitance and capacity and we monitor at the same time the capacitance right and so what we can see here for this material at about 4.6 volts or so we see a step change in the capacitance which we can also see in an 85 85-10-01 material however it happens earlier meaning it already happens at the lower potential and the reason for this is of course that we know that at 80 percent SOC we have the oxygen evolution but this happens at different potentials for these different materials so if we now put everything on the SOC scale in the excesses then we can see that these two plots essentially superimpose and relate of course very nicely with the oxygen evolution which we observe by mass spectrometry from these two materials so the oxygen release leads to a cracking or enhanced cracking of the material and increases the surface area due to particle cracking the other thing is when we use the single crystal material so this has the same composition these two materials the PC85 is polycrystalline the SC85 is single crystalline what we see there is that even when we go into oxygen evolution we see no particle cracking right and this of course also supports this idea that particle cracking mostly happens at the interfaces between the single crystallites in the polycrystalline material and since you don't have this single crystallite interfaces in the single crystals the surface area is quite constant and this is of course with the long-term stability which has been reported in full cells by the group of cheftan okay so then the next part we will look at the analyzing some of these degradation mechanisms by using x-ray powder diffraction and when we do this we first have to ask the question sort of what are the different mechanisms now so we of course already talked a little bit about it but essentially this is a cross section of an NCM polycrystalline material and for a cyclic material what you can see is cracks around all these primary crystallites and so what we expect is that wherever electrolyte has access we will form a reconstructed oxygen depleted surface layer on the on the NCM active material and that should lead to two things an increase in over potential plus also a loss of active material if the surface area if the surface layer is inactive for lithium intercalation anti-intercalation the other thing is of course we know that we can get particle cracking so it can become extreme so that really single particles could be released and then we would have a possible over potential increase because we have electronically now isolated particles sort of in the center of the secondary genres and the third one is of course the little mixing which in principle also should lead to an increase in the over potential due to the reduced availability and the reduced availability for lithium sites for intercalation so two effects also so of course what we find is that all these three contributions of course may play a role and we have in principle two effects over potentials an increase of over potential which can of course be caused by all of those which would lead to a smaller SOC window and thus to a loss of capacitance a capacity and of course we can also have an irreversible loss of material either by converting the surface into an inactive material by electronically isolating materials in the center of the secondary agglomerates or by having just less lithium sites available due to blockage by transition metal and so this capacity loss we want to examine by x-ray powder diffraction and the approach for this we set up a study where we looked at the long-term cycling stability of NCM811 at 45 degrees the C-ray was C over 2 at the cathode potential of 4.54 so this is of course a rather high cathode potential so it was supposed to be an accelerated study so what we used were pouch cells where we had a lithium reference electrode and where the graphite electrode was pre-lithiated so to avoid any sort of contribution from active loss of lithium and where the working electrode the NCM811 electrode was cited against the potential of the reference electrode and so we had six cells in this case which we cycled to different cutoff cycle numbers so six cycles 100, 250, 400, 550, and 700 and what you can see is that the capacity fading of these six cells follows each other very nicely here at C over 2 and here at diagnostic cycles at C over 10 the same is true for the voltage fading so when you get the decrease in the average discharge voltage and an increase in the average charge voltage due to the buildup of impedances and if we look at the capacity loss for this experiment over 700 cycles at 45 degrees at this rather high cutoff potential we lose about 68 million hours per gram which corresponds to about 70 percent capacity retention roughly and so then what we did is we harvested the cathodicative materials after these different numbers of cycles and in the discharge states and measured them in a capillary for x-ray powder diffraction and then in the charge state we used these electrodes charged them in a half cell and then repeated the experiment in the capillary and so with this we want to investigate the fading mechanism so this is shown in the next slide where we have a calibration curve for the C over A lattice parameter ratio as a function of the lithium content of the cam so this is in the fully discharged state this is in the fully charged state and what we find is this sort of calibration curve of C over A and from the C over A calibration we can get the amount of lithium which is in the material of course you could do the same thing by let's say ICP analysis and so these are the two in the two regimes in the discharged electrodes and in the charged electrodes you can get the relationship between the COA parameter and the lithium content and what we see is when we cycle the cells from beginning of tests to end of tests over 700 cycles in both the charge state and the discharge state we are narrowing the SOC window and from that we can calculate the capacity which would be due to simply cycling between these two windows between in the discharge state and in the charge state and so this delta lithium content multiplied with the total theoretical amount of 274 million miles per gram and then the 1.01 accounts for this 1 percent lithium excess we can calculate the capacity which we would expect based on these XRD measurements right and so what we record is the lithium content in the discharge state and in the charge state and here we can get the the shrinkage of the SOC window we can describe as a loss of capacitance due to over potentials which is essentially the capacitance which we get at beginning of test minus that at the end of test calculated by this equation here so when we plot this we can look at the capacity loss of course first the one we measure simply electrochemistry that's this part here and then the capacity loss which we would predict based on the shrinking SOC window detected by XRD right and so the difference between those two of course means that there must be a loss of cathode active material and we will try to a few slides later to quantify what this loss would be and so the other thing of course we looked at was was other possible capacity fading contributions one effect of course comes from lithium nickel mixing which we already discussed would induce higher over potentials and from a midfield analysis which was conducted analogously to what we had done in a previous study we looked at the amount of lithium in the of nickel in the lithium layer and what we see over these 700 cycles we get an increase of about two percent to be honest we expected the much more significant changes so once we consider this rather small so if you look at it what would this mean in terms of active material loss where you essentially blocking lithium sites that would be very little about four million miles per gram but of course what we cannot determine from here is what kind of over potential this would induce so it is a possible but most likely a minor degradation mechanism the other thing is the cracking of the materials we've seen the material good crack of a cycling and it is possible that you get the electrically disconnected materials inside the secondary agglomerate where we have poor electron conduction but of course a very good lithium ion conduction through the electrolyte and the cracks and this would result in the material loss from these isolated particles however if that were to happen what you would expect that in somewhere in the charge or in the discharge that you would have to see materials which have different lithium content so layered materials which has different lithium content and so for that we recorded the x-ray powder diffraction after all these cycles in both the discharge and the charge that this is just an example for the charge that you cannot really see much here but I can assure you we never saw any two different phases in lithium content right so from that we conclude that we do not really have a material loss due to isolated particles but of course we could still have a slower electron transport path here and induce an over potential so what we want to look at this well can we one part of the over potentials of course we can quantify that's the charge transfer resistance of the cathode active material and this we measured by two ways one was by a gcir pulse versus the lithium reference electrode so in principle you measure mostly the cathode response and what we can see here is that the impedance increases quite dramatically over these 700 cycles if we do a experiment where we harvest the electrodes and we build them in a coin cell with the freestanding graphite electrode so this is an electrode which has very very low impedance so this is shown actually here so the total impedance contribution of the freestanding graphite would be half of what you see here so negligible then we can analyze the this is a contact resistance on the cathode active material and then this is the charge transfer resistance so that one also increases and if we put it in the upper plot what we can see is that more of this follows the DCI resistance which really is what you would expect but what it shows is that the cathode impedance increases mostly due to an increasing charge transfer resistance which we believe is due to the formation of a oxygen depleted surface layer on the cap okay so now last slide what is the active material loss how can we actually determine it and so we can look what is our capacity loss from beginning of tests until a given cycle from cycle six in this case to 700 cycles we lose about 68 million miles per gram now we can calculate how much we would lose because of the shrinkage of the SOC window which is due to polarization due to over potentials that is calculated from the xpd data shown before and there we would lose about 31 of these 68 million miles per gram and then we can calculate what is the material loss this you could calculate from the difference of the capacity based on the SOC window minus the electrochemical capacitance normalized by the relative utilization of the materials so the details are really in the paper and that would come out as 40 now of course those two would have have to add up to this and as a matter of fact it's pretty good right so the sum of those two is 71 this is 68 so we have an error for about 5 percent the other thing we can calculate here is what is the percentage material loss so what fraction of our cam do we really lose and this would be calculate could be calculated quite simply for a given cycle this is what you calculate based on the SOC window minus what you measure electrochemically and this is also consistent what you would measure at a very very slow rate um or at the rate test going to very slow rates also detailed in the paper and so we can calculate the relative material last year and this is about 18 percent so that means over the 700 cycles at 45 degrees we lose about 18 percent of our hazard active material presumably due to the formation of a surface layer if that were true the estimated thickness of the surface layer sort of like a core shell particle if you want based on the BET area of a reasonably of a crack material as we measured before would be about 30 nanometer so this is reasonably consistent what people find in the literature if we plot this here this is the relative material loss this would be the calculated surface layer thickness over the 700 cycles we see in sort of continuously increases and if we compare this to a previous data we had done with exactly the same material using the same process but at much lower temperature then you see the higher temperatures actually enhances this effect and so what we can see is that we have a formation of an inactive reconstructed surface face that is more pronounced at the higher temperature and with this I'm a little bit late but I'm at the end so just to briefly summarize a few key points so we can monitor actually and see to the capacitance of an NCM electrodes and use it to estimate the extent of particle cracking and what we see of course it increases recycling and up on oxygen release at about 80 percent SSE but that only happens for polycrystalline materials and not for single crystals reinforcing this assumption that the weakest point is really between the single crystallites in a primary agglomerate in a secondary agglomerate the other thing is we see a surface reconstruction of an oxygen release which leads to an active material loss and an increased charge transfer resistance and from all we can tell is that the extent of lithium nickel mixing even at 45 degrees and 700 cycles is really not very large and so what they try to show is some diagnostic methods to try to decrypt with this capacity losses when cycling a cathodic material at some conditions so these were pretty extreme 45 degrees and a high cathode potential and what we see is that about 50 percent is due to impedance gain and about 50 percent is due to materials and with this I'd like to think our sponsors so really most of the contributions most of the data shown in this are actually through a project which we have had with BSF over the last 10 years and then some parts of it are funded by the German Ministry of Education and Research and this part was also funded by BMW and then of course I'd also like to acknowledge my group who has of course done all the work and even helped me make these slides and I thank you very much for your attention and I apologize for being a little bit longer than intended. Uber, thank you very much for the deep dive from degradation and diagnostics. We have a couple of minutes for a question before we move on to Young Cook. So maybe let's get started. I have a question if I may start first. What is the relationship between the material loss, the impedance and the impedance scores? So you talk about the BET surface area change at impedance scores but how about the the cracking part? Are they related or this is a bit separated? Material loss and impedance scores. Right, so I mean the material loss we are pretty sure is due to the formation of this inactive phase which we know also will lead to an increase in impedance. However, to be fair from the measured transfer resistances we could describe approximately half of these capacity losses due to impedance from what we measure. So the other half we believe must be due but we cannot quantify either due to the small increase in the lithium mixing because of course it is difficult to say okay two percent more nickel on the lithium sites what effect does that have on the resistance and end or due to you know higher electronic resistance into the inner part of the particle. So I think a better idea one will get with these studies with single crystals because then at least you can ignore the cracking part right this will not happen and however the lithium mixing part will still be there. Absolutely agree Hubert. Maybe a quick follow-up on that one so what always has confused me a little bit you know in terms of the BET surface area change and the impedance growth it's it's it's a little bit counterintuitive right so through the cracking in the delitiated states you're increasing the surface area but yet the impedance is decreasing and certainly many yourself included have pointed out the electrolyte reaction with the new surface but naively I also thought nonetheless you're increasing the surface area. Can you speak to why I would not expect the increased surface area to drop the polarization resistance in other words if I plotted the BET normalized charge transfer resistance it's growing very quickly right even more so than the unnormalized value. Yes I mean you're fully right right you would expect that as you create more surface area you have a lower polarization which I believe is also true the only thing is that most of the surface area gain from these very low 0.3 meters per gram to let's say 1.2 1.3 happens within the first cycle then it sort of gradually increases but then you don't have the factor between 0.3 and 3 anymore but you have the factor between I don't know 1.2 1.5 and 3 right so so the I mean this to be honest is I believe the reason why the rate capability of NCM is reasonably high despite the fact that they have as a powder a very low BET surface area because the effective BET surface area after one cycle let's say after the formation and that's when our experiments really started is already much higher so most of the benefit of the higher surface area due to cracking you get in the very first few cycles. Agreed yeah so I think the time scale of the resistance growth is also two-fold first cycle and then later cycle absolutely great very much so we have a question from our audience regarding the electrolyte chemistry so you discuss the transitive resistance growth can you maybe comment a little bit on how this would depend on the electrolyte choices. So let's say the formation of this oxygen depleted surface layer we believe does not depend on the electrolyte choice it's mostly a function of the state of charge to which you polarize the materials but of course it is true that when you have a let's say industrial electrode a thick electrode the state of charge is not necessarily homogeneous across the electrode particularly at higher rates so that unintentionally some of your particles near the interface between the electrode and the separator may already be at a much higher state of charge than your average state of charge and so in that sense it does relate to the electrolyte meaning to the transference number of the electrolyte and the conductivity of the electrolyte but I think it's only secondary. Understood Hubert so although we are out of time we did have one final question from Yang-Shan Horne at MIT so I thought I would ask it and I would just read her question for Baydem Yang asks could Hubert comment more on how the cracking occurs. How the cracking occurs so what we believe is so what we see for example for the single crystal material right which has a BET area of about one meters per program so this corresponds to about 400 nanometers or so 500 nanometers we see no cracking even though we cycle it in the same SOC range even though we go to very high SOC in one of the experiments right and so we believe that the cracking that the intrinsic strength of the material is high enough to accommodate the volume change which you do have without cracking and that the cracking only occurs at the interfaces between the primary crystallites. I think it was from Jeff Daan's group that I believe there were some studies on this but essentially this is sort of the weakest point right because you have these crystallites which move independently and the junction in between this is the weakest part and this is what cracks otherwise in single crystals you would have to see the same cracking. Thank you very much Hubert and we'll return back to you for a panel discussion after Yom Kook's presentation so let me hand things off to Yi. Yeah thank you Hubert for the very nice talk. Very deep dive on to the the castle materials. Let me now invite Professor Yom Kook's son to the stage. Yom Kook is currently a professor of energy engineering in the Han Yang University in Seoul South Korea. He received his PhD from Seoul National University in 1992 and then he was a later group leader in Samsung at the once institute of technology and contributed to the commercialization of the lithium polymer batteries. We all know what Yom Kook has been very actively involved in the research of all type of lithium related chemistry whether it's lithium ion, lithium sulfur, lithium air and also sodium ion batteries as well. To highlight one of his major achievements is the design of a new concept of layer concentration gradient MCM cathode materials for the lithium ion batteries. Yom Kook has published many many papers it's probably around the neighborhood of 500 papers right now. With that introduction let me invite Yom Kook to start his presentation. Okay thank you for introduction. It is my big pleasure to present my data and first of all I would like to thank the professor and professor to invite me to this very prestigious compulsion podium and I would like to talk today about the high energy cathode for next generation electric vehicles. This is the, this shows, let me see, this shows variation over the earth surface temperature between 18, 18 and the present and you can see the current earth temperature increased 1.29 degrees Celsius compared to late 19th century and it still continues to increase. We suffered from the hot summer in the last year to suppress earth surface temperature increase. One of the best solutions is why the spread use of the electric vehicles. This is the development history of the electric vehicles. At this moment we still use the generation two electric vehicles with the driving range between around 300 and 400 kilometers. To further increase the driving range more than 500 kilometers we should develop high-capacity electrode especially nickel-rich NCM and NCA cathode materials. This is the map of energy density of the cylindrical 1860-50 lithium-ion batteries. After the first generation of the lithium-ion battery was introduced in the market by Sony in 1991 the energy density of the lithium-ion has been increased three or three to four times. For example, as you can see here the great metric energy density increased around three times. The volumetric energy density increased four times. The optimization of cell design to maximize packing density is mainly contributed to such increase in the energy density. Of course, the modification of the negative and positive material partially contributed to. In order to further increase energy density we need high-capacity electrode especially in the cathode materials. This is my content, today's content. I would like to introduce capacitive fading mechanism of the nickel-rich cathode first. And this is the charge to charge codes. As you can see that with the increasing nickel content this charge capacity almost linearly increased. And the surprising that you can see that here the pure nickel oxide material delivered the discharge capacity of around 250 mA of currents. On the other hand the cycling performance is decreased with increasing nickel content. And as you can see from the DSC profiles with increasing the nickel content exothermic peaks shifted to lower temperature with higher heat generations. This is the summary of this start. With increasing the specific capacity by raising nickel content from one third to 100%, thermal stability and capacity retention is accordingly decreased. However, based on this result, we believe that it is impossible to develop ideal cathodes with both high capacity and high safety just by changing the compositions. However, we need the cathode material at this point with high capacity, good thermal stability and outstanding capacity retention. In order to develop the target cathode, we should know the capacitive fading mechanism for the nickel-rich NCM and the NCA cathode material. This is the DQ-DV curves and the volume variation of various NCM cathode materials. As you can see that NCM 90 and 95 and pure LNO exhibit four distinct redox peaks due to multi-phase reactions from the H1 and the monoclinic H2 and H3. And the redox peaks became polarized and reduced in the height with the cycling, especially in the H2 and H3 phase transitions. On the contrary, these redox peaks of the 622 and 811 hardly changed during the cycling. As you can see, the unicell volume variation charging voltage, the unicell volume decreased monotonously up to 4.1 volt and then decreased rapidly above 4.2 volt, corresponding to a pex of the H2, H3 phase transitions. The unicell volume variation decreased with increasing nickel content. For example, this value is 4.3% for 622 and almost 10% for the pure LNO particles. And these are the cross-sectional SCM images charged to 4.3 volt at the first cycles. And as you can see that over the microcrack in the cycles 622 and 811 NCMs were rested before reaching to particle autosurface. However, the NCM95 and LNO increased the amount of the microcrack which propagated to particle autosurface. This is the degradation mechanism of the NCM cathode. In the case of the nickel content, more than 80%, the formed microcrack resulted from the H2, H3 phase transition propagated to particle autosurface, facilitating electrolyte infiltration along the grain boundary into the particle interior, which exhalate surface degradation of bromide particles by reacting unstable tetravalent nickel with electrolyte to form nickel oxide like impurity layer, which leads to gradual capacity fading. In order to overcome the intrinsic property of the capacity fading of the nickel-rich NCM and the NCA cathode, we developed two approaches for the last 20 years. One approach is concentric gradient, the other approach is microstruct control cathode. Let me introduce concentric gradient cathode. This is the development history of the concentric gradient cathode materials. In 2005, we reported coaxial material called generation 1. As you can see, this is the concentration profile of transition metals. In the case of nickel concentration, concentration is kept at very high levels and then suddenly decreased at the particle surface. In 2005, we developed coaxial with concentric gradient materials, and in 2012, we developed full concentric gradient material, nickel concentration profile like this, and manganese and cobalt concentration profile like this one. And three years later, we developed advanced full concentric gradient with two slabs. This is the schematic diagram of the synthesis of the concentric gradient hydroxide by the core precipitation, and we can use the same facility in synthesizing conventional hydroxide particles. Additional facilities is one solution reservoir tank. As you can see that in these figures, nickel-poor solution in tank 2 is slowly pumped into nickel-rich solution in tank 1, where the mixed solution is fed into batch-type reactors, leading to smooth concentration gradient of transition metals within the particles or during the core precipitation process. Concentric gradient cathode materials have unique pictures distinguished from the conventional cathode acting material. First, as you know, as I explained, concentric gradient cathode material consists of nickel-rich core and nickel-poor shell part. As you know, as you know that the cathode surface composition comprising high nickel content is very reactive to electrolyte attacks. The nickel-rich conventional cathode surface is easily damaged by the electrolyte attacks forming the thick impurity layer, where is the nickel-poor surface of the concentric gradient, the cathode is much stable electrolyte, much stable from the electrolyte attacks, as you can see here. And the secondary concentric gradient cathode consists of long rod-shaped primary particles, where this conventional cathode is composed of large equi-axid polygonal type shaped primary particles. As you can see the cross-sectional SM image, the conventional cathode material easily developed significant micro-crate induced by the internal stress during the charge. However, the rod-shaped primary particle of concentric gradient can effectively dissipate internal stress and those minimize micro-crate formation within the cathode particles. This is one example of the concentric gradient. We synthesized two cathodes. One is an FCG full concentric gradient with a nickel content of 61. The other one is 1 percent aluminum of the FCG material called AL FCG and then we compared long-term cycling performance of the two cathodes. This is a microstructure of the full concentric gradient material. As you can see the EPMM mapping image, the nickel was depleted at the particle surface and became gradually increased toward the particle centers. EPMM line scan also verified successful synthesis of the FCG cathode active materials. As you can see from the EPMM images, surprisingly FCG cathode particle composed of the long rod-shaped primary particle aligned toward the particle centers. Their lengths estimate to be around 2.5 micrometers. Another unique picture of the FCG material is that all of the observed primary particles have their C-axis aligned in normal to AB plane, providing the best channel for the region diffuse. This is the comparison of the cycling performance of commercialized NCA-82 together with the aluminum FCG cathode materials. The long-term cycling performance clearly demonstrates the superior region intercalation stability of the aluminum FCG cathode. As you can see from the same image, nearly all over the particle from the cycled NCA cathode was completely publicized. In comparison, in the cycled FCG cathode, the original spherical morphology was well preserved. In addition, the composition line scan confirms that the original concentration gradient was well maintained even after 3,000 cycles. We did a safe test. One is the nape penetration. The other is an overcharged test. We fabricated pouch-type cells in our laboratory using the synthesized FCG cathode with the capacity of 250mAh, and then fully charged to 4.2V, and then the nail was penetrated. After the nail penetration test, the highest cell temperature over the cell was 70 degrees Celsius. We also did an overcharged test. The overcharged test was carried out, charged to either 250 SOC or 12V. As you can see, the pouch cell was slowly swollen due to electrolyte evaporations. As you can see here, after the overcharged test, the cell voltage increased only 5.5V with the temperature remaining below 20 degrees Celsius. After the nail and the overcharged test, both cells showed no smoke, no thermal rays. This is another example of the concentration gradient. We prepared two NCM-90 cathodes. One is the CSG-90. The other is a conventional cathode without a concentration gradient called the CC-90. Then we compared the structural and electrochemical performance of the two cathodes. As you can see from the long-term cycling performance, the CSG cell shows much improved cycling performance with the capacity tension of 80-80% after 1,000 cycles compared to 68% for the CC-90 cathodes. As you can see from the closer sectional SM image, after the 500 cycles, the microcrack and the particle fracture was observed in the CC-90 cathodes. Furthermore, after 1,000 cycles, the CC-90 secondary particles nearly pulverized into the individual primary particles. On the other hand, in the case of the CSG-90 cathodes, no visible microcrack was observed after 500 cycles and only hairline cracks, hairline microcracks were observed within the particle interior after 1,000 cycles. In order to understand the phase evolution, the convoluted 003 reflection of the H2-H3 phase as a function of state of charge are studied. As you can see, the 003 reflection for the CC-90 cathode shows the whole existence of the H2-H3 phase is only detected around only 4.2 volts, indicating sharp phase transition from the H2 and H3. In the case of the CSG-90, the H3 phase began to appear above 4.2 volts, but the H2 phase was observed up to 4.3 volts. Moreover, the CSG-90 cathodes less support from the volume variation compared to the CC-90. This is a 10-image of the two cathodes after 500 cycles. CSG-90 cathode particles, egotivity, thin nickel oxide-like impurity layer of 5 nanometers on the particle surface. On the other hand, surface damage of the CC-90 is more severe than the death of the CSG cathode because of the nickel-rich outer surface showing the impurity layer of 30 nanometers. In addition to surface damage, the interior primary particle also suffers from severe surface damage due to electrolyte attack through the formed micro cracks. To confirm the outstanding mechanical stability of the CSG cathode, stress distribution was calculated. This is the CC-90 cathode. This is the CSG-90 cathode. In calculating shell, compress the core and lead to slightly smaller tensile stress in the core than in the CC-90 particle to reduce tensile stress, suppress the micro crack formation, and thus improve the mechanical stability. As you can see, more importantly, how to share egotivity, much more homogeneous stress field compared to CC-90 cathode, which suppress crack growth inside the shell. This is the characteristic dating mechanism for two cathodes. Based on this result, we conclude that outer shell is very effective in suppressing micro crack growth within the shells. Our concentration gradient cathode material already penetrated into the EV market. This technology was licensed to three Korean companies. In 2018, Gia-Niro EV used these materials in 2020 in the Hyundai Kona EU and AcaFox Micro 5 from the Beijing Motor Corporation also used the concentration gradient cathode materials. And we are expecting the concentration gradient cathode material to enter into EV markets. I will briefly introduce microstructure controlled cathodes. Inspired by the concentration gradient cathode materials, we modified the nickel-rich cathode material without the concentration gradient. The conventional cathode, conventional NCA and NCM cathodes were composed of randomly oriented polygonal shaped primary particles. And microstructure can be modified by the X-doping. The modified cathode consists of a radially oriented, large-shaped primary particle as you can see here. Let me show our typical example. First, we developed all-on-doped nickel-rich NCM cathodes. We prepared two cathodes. One is pristine NCA, the other is one more percent all-on-doped NCA called BNCA. And as shown in this figures, the microstructure of the BNCA cathode material notably changed to have a long, large shape, prime particle with the length of the micrometer. Accordingly, the BNCA cathode shows much improved cycling stability with capacity tension of 83 percent after 1,000 cycles. And as you can see from the cross-sectional HCM image, after 1,000 cycles, PNCA cathodes of particles were nearly pulverized into several segments, whereas PNCA cathodes maintained their original particle integratives. The superior cycling stability of the BNCA was further confirmed by the in-situ XRD measurement before and after 1,000 cycles. As you can see there, the PNCA cathodes demonstrate smooth-based transitions after long-term cycling, whereas the H3H is not observed in pristine PNCA cathodes after 1,000 cycles. In addition, the change in the change in contour plot of the 003 reflection agree well with k of the H2H3 phase peak intensity. The H2H3 redox peaks for PNCA cathodes disappeared almost completely after 1,000 cycles, while PNCA cathodes maintained the distinct redox peaks after the same cycling period. To estimate the extent of the surface damage over the surface damage on cycling, nickel oxidation states were mapped by the typographic soft X-ray combined with the X-ray absorption spectroscopic. As you can see that in pristine PNCA cathodes nearly all over the interparticle boundary was completely exacerbated by nickel 2 plus phase. In comparison, although some region over the PNCA suffered from the surface degradation along the grain boundary, the distribution over the nickel 3 plus is uniform, confirming enhanced cycling stability over the PNCA cathodes. After verifying the microstructure modification effect, we further investigated to find the optimum microstructure for achieving long-term cycling stability. This figure shows the dopant effect on microstructure over the cathode primary particle. As you can see the microstructure varied from large to actually polygonal type primary particle to fine needle-like particles. The cycling stability as you can see is strongly dependent on the particle microstructure. Among the various dopants tested, the tantalum of the cathode material shows the best cycling performance. This is the cross-sectional SM image over the NCA and the NCPA cathode material. In the case of the NCA, the micro-cracking becomes severe with increasing cutoff potential and the particle was correct into the several segments after charging to 4.3 volts. In comparison, NCA cathode contains fine microstructure which are arrested within the particle cores. We fabricate electrode by mixing over NCA and NCPA cathode material and then charge to 4.3 volts. Cross-sectional SM image clearly demonstrates the superior mechanical stability over the NCPA cathode materials. Moreover, the aerial correction over the micro-crack in the NCA cathode during the discharge is larger than those during the charging at the same voltage, suggesting that micro-crack do not completely reversibly close. However, NCPA cathode demonstrates reversible micro-cracking, opening, and closing behavior. This is a mechanism that enables superior cycling stability over NCPA cathode materials. As you can see that compared to other NCX cathodes, Tantallium dophid cathode is the best long-term cycling performance and cross-sectional SM image confirms morphological integrity of the NCPA cathode materials. We found that microstructure for primary particle depends on the open and conclude that there is optimum microstructure such as aspect ratio and primary particle risk for achieving long-term cycling performance. This is the high-resolution and hard-off time image of the NCPA cathode materials. As indicated by the red arrow, we observed an additional extra spot in the layered structure patterns. In addition, as shown in the high-resolution PM image, some regions showing the extra spot is also observed in the layered structure. As shown in these extra spots come from long-range order, the interchange of the region and nickel ions, so-called ordering structures. We believed that the interchange between region and transition metal induced by high-balance Tantallium dophid, which changed charge distribution around itself. This structure seems to enhance structural stability by suppressing interlayer collapse at the highly-charged state, and we are further studying these ordering structures. This is my conclusion. It is impossible to develop ideal NCHM and NCHCathode just by changing compositions. Unlike NCHCathode, aluminum F3G was cycled at 100% DOD for 3,000. All on NCHCathode greatly improved cycling stability. The Tantallium substitute cathode produced readily oriented primary particles. The superior cycling stability clearly indicates the importance of microstructure. We believe that our strategy of optimization of the cathode microstructure can lead to rational design and development of NCHM and NCHCathode. We thank the Korean government, BMW, LG Energy Solution, BHSF, or CDMM for supporting this research. Thank you for your attention. Well, thank you, Yangku, for the very interesting result. Let me ask you a couple of questions. I think we have a little bit of time. The first one, Yangku, you show this very nice morphology control, whether you go from polyhedral shape to this more like needle, nano-rot shape, kind of spiking going on forming these secondary particles. What's the crystallography orientation if you have this particle, this needle shape, pointing out and what's the A and B, C axis, right? Particularly along the length of the rod. What's the axis? The reason to ask this question is, the volume change, you show the unicell, when you take lithium out during deletion, the unicell shrink, then looking at A, B, C lattice constant, and which one shrink the most, then this morphology matching with the crystallography orientation, whether it is a correlation, how you arrange in these spherical particles that can help avoid the cracking. I'm trying to establish that correlation. Want to see your thought on this? Yeah, that is a great question. As you can see, this is the high-resolution PM image. The large shape morphology has the A, B orientation in this direction, and the C axis is aligned in these directions. And therefore, lithium can be easily intercalated or deintercalated through the primary particle, through the A, towards the A, B planes. In addition, as I shown before, and let me see, yeah, in this case, as you can see, stress distribution calculations, as you can see, this is the C directions. During the charge, she actually expanded at the highly rigid state. And however, the stress is uniformly contributed only to the shear regions. Therefore, we cannot overdue microcracking severely. Yeah, so very good. So looking at Hilbert's talk, my Hilbert has this diagnostic tool during charging, this charging, looking at impedance change, surface area change. So the different composition and morphology control, in your case, some of them has less cracking, like the full gradient one has less cracking, this needle shape has less cracking. Would you be able to do this impedance study yet, like what Hilbert just showed, to see during this process, how impedance will evolve, and the surface area will evolve with charging, and this is charging cycle for different morphology for particles you have, whether that correlate with the performance. Yeah, the Hilbert did a very nice research in the identified passive fading mechanism in terms of the B surface area and the cracking. And we did also almost a similar experiment. We checked the impedance variations during the charge, during charge and with the cycling. And then compared the conventional cathode with the large-shaped morphology cathode materials, based on our expected result, the impedance of the conventional cathode rapidly increased with the cyclings. However, our large-shaped morphology cathode material shows a very stable cycling, very stable charge stress for impedance increase, even though irrespective of the cycling, the charge transfer is very stable, which is much smaller than those of the conventional cathode materials. And we also checked the specific surface area changes during the cyclings, based on our result, the B-T surface area not so significantly increased with the cycling compared to conventional cathode materials. Okay, yeah, so let me ask you one last question, and then I'll bring Will Chu and also Hilbert to the stage again for panel discussion. From the audience, there's one question about, well, how does the dopant affect the necessary sintering conditions, such as temperature? If you choose different dopant, how does the sintering condition you will need to consider? And also, whether the effect of ammonium hydroxide content compared to the different dopant use, you know, what parameters have a really large influence in controlling the primary particle morphology? Yeah, that is also very nice questions. And we are now starting the research, how to control the primary particle morphology by doping and other technologies, other things. And based on our recent result, some dopant is very effective to prevent sintering of the primary particles. In this presentation, I didn't explain the morphology of the hydroxide prickles. In our hydroxide prickles, hydroxide prickles have long rod-shaped primary particles. And if the hydroxide prickles have long rod-shaped primary particles, and if we add some dopant, the rod-shaped morphologies will preserve even after the high temperature concentrations. We are understanding intensively why these dopants prevent the sintering effect. Yeah, that's great. It's good to know that dopant has such a big effect and stabilizes the morphology during sintering. So with this, thank you, Yankut. Let me now bring Hilbert and Will back to the stage. So this is for a panel discussion, you know, this certainly questions freely flow. So maybe the first one, Hilbert, I'd like to pick your thought a little bit. You have seen Yankut's talk and Yankut see your talk. I'll give the opportunity for both of you to mutually, maybe you have one question to ask. I see there's a lot of synergy between your two talks. Do you want to mutually ask each other a question? Yeah, I mean, maybe an observation from Yankut's talk, and just to understand whether I got this correctly, but from the analysis I saw, the conclusion is in your case also, Yankut, note that the particle breakage always occurs at the interfaces between the crystallites, right, and that you can affect this by how you arrange crystallites in a secondary agglomerate, whether you do it like this or, you know, in a spherical manner. But it always occurs at the interface and not through a single primary crystallite. Is that correct? Did I see this correctly from your talk? Actually, I don't know exactly whether the primary particle is single crystal or not. I think this is, that is dependent on the nickel compositions. We checked the crystal structure of primary particles within the secondary part of the nickel content, lower nickel content, cancel electron material. For example, nickel content was 60%. We confirmed that the primary particle was single crystallite. However, we didn't check the crystal structure in detail over the primary particle in high nickel content, such as 90%, because as you can see that primary particle morphology over the high nickel content is quite different from that of lower nickel content in terms of the particle shape and the particle thickness, the particle length. And that is our future homework to identify crystal structure and so on. And then the other question I had was when you look at the effect of dopants, right, which then result in very different materials at the end, is the effect of dopant to change the precursor material morphology, which then carries over into the morphology of the calcine material, or is the effect of the dopant affecting the way different morphologies are being formed during calcination? Yeah, actually, we are preparing the papers concerning this point. And based on our results, I think the hydroxide precus microstructure is very, very important. I mean that without the primary particle within the hydroxide precus, we cannot make large-shaped columnar-structured calcine material after the calcinations. For example, we prepared two different kinds of hydroxide precus. One has the long, large-shaped primary particle. The other is the conventional hydroxide precus without meaning that without large-shaped morphologies. After that, we doped the same dopant such as for example, Tantanu. And in some case, the cathodic material shows the columnar structure, large-shaped columnar structure or line structure. However, some hydroxide precus cannot make large-shaped morphologies. Based on this result, it is necessary to make two synthesized hydroxide precus with large-shaped morphologies. That's great. Yang, could you have a question for Hubert, or should I pass to Will, for Will to ask you both of your questions? Yeah, Hubert, you did a very scientific and very nice analysis in terms of the in-situ impedance and so on, in-situ XRD. As you know, there are a lot of variable for the capacitive fading for the cathodic active material, as well as such as changing of the surface area and the surface changing of the crystal structure and the change of the microcracking. As you presented, the capacitive fading is mainly dependent on the three variables, microcracking, cation mixing, and increased surface areas. However, in the case of microcracking, some isolated particles cannot contribute to the impact, such as impedance and XRD. How can you differentiate among these parameters? I tried to briefly explain it, but very briefly only. Essentially, when we have significant particle cracking, you can imagine that particles inside may be electronically very poorly connected or disconnected. Some people talk about these so-called fatigue phases or something like this. In that case, however, you would have to see it in the diffractogram because you would have to see layered materials of different lithium composition, which is quite easy to distinguish. That we never saw. From that, we concluded that we didn't have perfectly isolated material. However, we could not exclude that you may not have a higher polarization impedance because of this. That cannot be excluded, but you can exclude that you have completely disconnected particles. That's for sure, because if there were, you would have to see it in the XRD. Overall, it's of course very difficult to completely quantify the different contributions. The only contribution one I think can quantify is the material loss you have due to the reconstruction of the surface and the impedance of the charge transfer impedance, but the contributions by, let's say, lithium nickel mixing, which does increase over time, but still remains at a very low level. We cannot put into numbers or quantify what defect it would have on the capacitance or capacity. Sorry. That we could not determine. I mean, what we had hoped for, to be honest, because we had done this study at 25 degrees, and we didn't see any significant lithium nickel mixing. What someone had suggested was like, well, why don't you see the entire temperature because there should be some significant mixing. And so that's why we conducted the study, but we didn't see anything, unfortunately. I mean, it was the same level as we saw in room temperature. You mentioned that the capacity loss is partly coming from the loss of the cathodicte material. Did you measure this amount from the anode side, graphite anode side, or rich metal side? I think because I think the capacity, the mass loss is coming from the isolated particles within the particle centers. Well, as I said, I think we excluded the isolated particle hypothesis because we would have expected that we would see it in the X or D, either in the discharge or the charge state, because they would have to have a different lithium content. So that we didn't see. Now, any effects of the anode, and this study are completely eliminated, because we have LTO anode, which was pre-lithiated. We have a lot of electrolyte and so on. So this study really was just a simply only focus on the cathodicte material. And then the amount of particle loss, yes, I think you can get that pretty conclusively, because you know the SOC window in which you cycle, so you know what capacitance that should give you. And we know the electrochemical capacitance we measure, and the difference is the material loss. And you get the same result instead of doing the, I mean the X or D analysis is nice, but this very cumbersome. And actually you can get the same results by doing very, very slow rate tests. So if you do a rate test from high rate to very low rates, you can extrapolate sort of your material loss. And that agrees with the X or D. And this is admittedly much simpler experiment. So material for sure is gone. And we believe we can exclude that it's isolated material. And let's say the amount of the material which you would grow as a surface space on the cathodicte materials is reasonably consistent with what people report in the literature in terms of thickness. But of course, to really truly convincingly demonstrate that one would have to do some detailed TEM measurements to measure those. I mean, that's for sure. This is very nice discussion. Where do you have a question you want to ask the panelists? Well, the moderators are almost not necessary here. So I think I have now found the recipe to have a great discussion is to have it, you know, Friday midnight or Friday afternoon. And really, the blood gets flowing a lot. So I'm really enjoying the discussion. Yankuk, you made a very important statement. It was the first line in your conclusion, which is, it is difficult, if not impossible, to simultaneously realize safety and capacity. And you made this point very clearly a number of years ago with the very famous plot, showing the oxygen release and the temperature and also the energy with the nickel content. So I would like to probe a little bit deeper to the both of you, Hubert and Yankuk, how as we go to these very, very nickel rich compositions, it seems that all the modification is only having a negligible effect on the oxygen stability of the system with regard to exothermic reactions and safety. So how do we reconcile this too? What is the strategy to getting the safety back in the system if there is a strategy? Or is this something that we have to do at the systems level, maybe at the battery pack level or thermal management to combat this issue? Maybe Yankuk can comment? Yes, Hubert mentioned that he studied a lot of, he studied evolution of the evolution mechanism of oxygen. He published many papers and based on the, I just focused on the cracking mechanism. And I think the cracking and the oxygen evolution occurred simultaneously because the, as you know, tetravalent nickel is very, very unstable at a highly charged state. And tetravalent reactive and unstable tetravalent nickel is changed, automatically changed to a more stable phase, such as the nickel oxide, the rock side phase. During this process, oxygen is automatically evolved from the host structure. And therefore, in order to prevent oxygen evolution from the host structure, meaning that other ways, meaning that the stabilization of the cathodic material by preventing cracking is, is that we should make, we should synthesize colorless structure of the cathodic material because even though the main capacity fading is coming from the inner particle, inner primary particle within the secondary particles, because the surface area of cathodic material are not so big, not so high. And however, if particle cracking is happened, as Hubert showed, the surface area increased, dramatically increased, which increased the plastic reaction with the electrolyte. And therefore, we should prevent micro-cracking to cracking. And thus, the morphological integrity should be, should be maintained. And I mean that other ways, if the micro-cracking is happens, in the exposed surface area within secondary particle is huge, and which induced nickel oxide, nickel oxide, the phage, nickel oxide, the phage formation, together with simultaneous, together with oxygen evolution. Can I maybe, before Hubert, you also share your thoughts, let me just ask Jungkook a quick question. Do you believe there is a way to delay the oxygen evolution, exothermic oxygen reaction on heating, right? So to really think about ways to improve the safety. So in other words, if you have a particle that doesn't crack at all, do you think you could substantially raise the oxygen release critical temperature? Yeah, I don't know exactly the question. Maybe Hubert knows better than me, because he studied oxygen evolution reaction intensively. To be honest, I would go back to your data, because you published this beautiful data, right? That showed very clearly, right, that the higher the nickel content, the lower is the temperature at which you release oxygen, you know, at about 200 degrees or something like this for a really nickel rich material, right? So that problem, I think, will not go away. So when you say safety, right, there's always a question, what safety? Is it over charge or is it overheating or whatever? But let's say overheating, that problem, I think, with the nickel rich material for sure would be there, right? Because, and it's kind of funny in history, because people had NCA and they said, well, NCA is a little bit unsafe. So then people came up with NCA. So they used the NCA-111, which was quite safe. But then they said, oh, but it doesn't have enough capacity. So they made the NCA-8505 or whatever, right? And now it's safety is the same, right? Because it's just a nickel content. So I think in terms of intrinsic safety, I mean, you, I mean, there is the high voltage spinel, right? That doesn't release oxygen, neither electrochemically, nor thermally. I'm just, you've got really high temperatures. And the other material is lithium and manganese, manganese rich material, right? So I think both would have intrinsically a much higher safety. Huber, I really resonate with your point. It is a trade-off, right, between performance and safety. And right now, this is really driven by market requirements. And, you know, it really depends on how much you value each. Exactly. But this also I think could be a segue for me. If Eve, you don't mind, I can ask one more question, which is, given all these trade-offs, what should the roadmap be for cathode chemistry? I think this is the, you know, literally the trillion dollar question that people are asking everywhere, academia, industry, and the like. You know, certainly we are already approaching almost completely lithium-nickel oxide. Surely there'll be a little bit of dopants and such, but we're getting very close to the maximum capacity. What's coming next? Maybe I can provoke the both of you to comment a little bit on this and try to forecast where the next generation cathode would be. I mean, there was a recent discussion with some OEMs, car manufacturers, right, to where it was like, well, you know, what about safety and do you compromise safety? And they said, we never compromise safety, right? And I think it is really true, right? You can control even this very reactive chemistry, but it is additional cost to your system, right? You have to have additional safety features in your system, you have to cool the battery, you have separate sensors and whatever, but it can be done, right? And so at the end, it's just the question, well, how cheap can it be and what is the cheapest system? And so of course, if you can make a material which is very safe, you can save a lot of systems costs, right? On the other hand, well, maybe you have a material which is not so safe, but very cheap, you can afford it on the other hand, right? So I think it's not the question one can look at in an isolated fashion, right? Because at the end, it's the cost of the entire system. And at the end, it is for sure true that I mean, this was a person from Volkswagen who said that we do not put cars on the road, which we do not consider safe, not in the millions, right? I mean, that would be crazy. And so I think, yeah, so I think it's been demonstrated that the safety of these materials can be maintained. It's just always a question at which cost, right? And so I think if you look at sort of the energy targets, people have, you know, for electric vehicles, 60 kilowatt hours is almost standard, right? And they want to go to 100. I mean, at the end, then it's cost, right? And then when you look at cost, then I think it's pretty clear you have to get rid of nickel because only manganese could meet the cost targets, right? And so I think the two chemistries which are intrinsically safer, namely the manganese rich chemistry, be it the spinel or be it, you know, the listed manganese rich materials, I mean, they of course would, if one can make them work to the extent that they meet all the lifetime requirements, they would, of course, hit both targets. So that's why I think it is quite important to look into these materials because I think this is the only way to realize cars with large batteries, right? Even larger. I mean, to a normal cost. Thank you, Hubert. All right, that's answered in the trillion-dollar question. Yanko? Yeah, I think there are two ways to develop the cathodic materials. One way is continuous increase in the nickel content to deliver high capacity. And we should make the good material as possible as possible as we can. And for example, as I told you, as I told you, as presented, and we should synthesize why we can should design nickel-rich material with the columnar structure without the micro-crackings, which increase the stability further as well as the cycling behavior. And if we make the best cathodic material with the nickel content, even though, even though, even though, which shows the poor thermal stability, cell company, cell make will optimize the rich of iron basedly by a combination of the end of the material and the electrolyte and the other materials. That is one way. And another way is to develop the manganese rich, rare cathodic materials. And as you know, rich rich and manganese rich material is the, even though one strong candidate, however, there are a lot of research for more than 20 years. Nobody's achieved in overcoming these issues. Another way is my opinion is that we should develop the manganese rich, manganese rich, layered materials. The question is how to stabilize the crystal structure, layered crystal structure for manganese crystal axiometry. And my question is that we should do stopping and making the, making the morphology such as columnar structure, something like that. And that we should, we should deeply investigate this way because this research was not intensively before. And the people say that manganese rich, we cannot make the, we cannot synthesize the layered manganese rich cathodic materials. And we didn't, we didn't intensively study this material. That is my, my, my opinion to develop good cathodic material. So, so I think our time is getting close. I have one last question, but I'm going to ask the question you guys don't need to answer our lab bill to wrap up the whole day and also for the audience to consider. So will all this discussion change, right, dramatically when we go into the solid state batteries regime? I know it's very early solid state batteries. Maybe it is hard to have a detailed discussion right now, but if you want to have, you know, 30 second each to say what you want to say about solid state, whether that will change the whole thinking. Hilbert looks like you will have something exciting to say. Oh, the solid state. Well, I mean, in general, right, I mean the solid state batteries still use the same cathodic materials, right? I mean, maybe slightly different morphologies or whatever, but the chemistry of the active materials is not any different, right? And I mean, one of the big advantages, of course, potentially would be that you could use a lithium metal anode, right? But it's not so straightforward either because the solid electrolytes are not so solid, right? I mean, dendrites can still go through. And, but yeah, so in terms of, you know, there may be some advantages in terms of temperature stability or so, but as far as the active material degradation per se by itself is concerned, I think you would have exactly the same, the same phenomenon, right? You would have the same oxygen release and you would have, if you have a diluted, a strongly diluted NCM, it would release oxygen if it gets warm, right? So that I think wouldn't change anything. I think it has many advantage, potential advantages, I think, but as far as just reducing it to the cathodic material, I think it's very similar. Yeah. Well, young code, 30 seconds. There are a lot, a large number of hurdles to develop the solid stability in terms of the cathode and electrolyte and anode material too. And we should develop, we should solve one by one step by step. And as a part of the cathode active material, cathode active material should requirement of the cathode active material for solid state battery is very, very resistive material to the high pressures. Because in order to make the solid, the good solid state battery, we should, we should have intensive flashing between electrolyte and the cathode active material. Within under this circumstance, the cathode active material is very resistive for the problem of pressure. And that is the cathode active material requirement. In addition, in addition, surface should be very stable from the reactive sulfide or highlight electrolyte. As for the electrolyte, we should develop the more stable, more reliable electrolyte rather than sulfide electrolyte. And we should, in order to mass production of the solid state battery, we should use the lithium ion battery facilities. In order to do that, we should develop the very stable, very stable solid electrolyte. And as for the lithium metals, lithium metals still has a tender light problem. Even we use the solid state electrolyte. And we have a lot of hurdles. And I turn to you. That's probably for another day. Well, we, we thank you, Yankou, and thank you, Hu, but we start real back back to you. Yeah, let me add my thanks for a very illuminating and provocative discussions, both technically and on a broader note. So thank you both again. And with this, we are wrapped up for the screen quarter seminar series. We will return after the fourth of July holiday in the United States with our summer series of seminar. So please stay tuned for our announcement on the next series of speakers. Ah, I see. We also have a rescheduled talk. So some of you might remember that Tim Holm from Quantum Escape was to speak in June, but was unable to. So we have rescheduled Tim to July 30th. So we'll announce the talks for the next quarter in the next couple of weeks. So with that, I'd like to thank everyone for tuning in this spring. And hope everyone will enjoy my mostly pandemic free summer and hope to see you in July. Thank you very much.