A lecture from Berkeley Lab's Environmental Energy Technologies Division covers some promising materials research efforts that are expected to lead to improved battery technology. Mark Verbrugge, the director of the Chemical Sciences and Materials Systems Lab at General Motors' Research & Development Center, discusses the research.
Transcript
00:00:00 thanks for coming out today it's a a pleasure and an honor to visit here on the hill and also uh various faculty members uh down at uh at Berkeley uh the university uh as John mentioned I'm going to talk about electrochemical energy storage Technologies and the automotive industry key drivers and needs um the outline of the talk is shown here I'm going to start out rather
00:00:20 General and talk about some of the issues that we face as Society in terms of energy sustainability much of this is probably familiar to you but it helps set the context for some of the subsequent work in the second portion of the talk then I'm going to go a bit more into what is a major Trend in automotive today that is lithiumion traction batteries uh you
00:00:39 may not know this but although there's a lot of work in this area there's actually no high volume lithium ion traction battery out on the road today it's one of the things that's going to become IM imminent probably about post 2010 um so I'll focus on uh lithium batteries look at some new results with my collaborators at H uh what used to be called hug research lab now called uh
00:00:59 hrl laborator that's half owned by Boe and half owned by General Motors and we have some battery projects there and then secondly some work I've done with a colleague at University of Kentucky on stress strain distributions in insertion electrode particles um and I'd like to also cover some of the promising develops in developments in lithium batteries and then I'll uh I'll draw
00:01:18 some summaries uh with regard to the the high level context um this gives a bit of what's going on both in terms of the major driver obviously world population and that's on the left ordinate then uh over here and uh you can see that uh in today's uh terms we're hovering between the uh 5 billion Mark and and plus or minus around that region and I put a
00:01:41 blip in here as far as the right side on the vehicle park vehicle park is all the vehicles that are out there not just those that are sold uh uh annually and the blip is because obviously we've gone through a rather significant financial uh turmoil in in the last year or so and so uh that's the reason for it uh one of the things that's clear actually on this chart is that this Trend continues a
00:02:01 automotive business is actually a very good business to be in it's continually growing in terms of the marketplace but B there's a question about sustainability um in terms of GM uh uh specifically obviously right now it's still going through some of the ShakeOut of this last turmoil um but uh as of two years ago we sold more Vehicles outside of uh United States than within the
00:02:21 United States and that was a change in trend for the industry and so it's the growth in the developing markets that's actually uh driving this uh industry in terms of vehicle volume um a plot that you'll see if you plot almost any commodity or any item that is purchased is a monotonic increase with the per capita income so here we're looking at energy consumption uh per
00:02:44 capita on the ordinate and then the annual GDP per capita income in US dollars in this case $2,000 and you can see this generally speaking monotonic income you're going to see a couple other plots that reflect this as well and again it drives back to that notion of sustainability uh these are 2005 data uh this is a rather recent publication by
00:03:02 the group out at MIT uh some of you may have uh seen some of their work in in recent times and it's it's actually quite enlightening if you look for example at this uh black curve shown here ranging from 1950 at the start to uh 2005 you can see the effect of GDP per capita uh on the uh epsa again and then that effect on the passenger miles traveled vehicle miles traveled on the
00:03:29 ordinate and basically people travel a whole lot more when they have a a higher income and then the asop that they're trying to infer sort of interesting here it's a little bit hard to tell as plotted but right around here is if everybody took a jet airplane to travel around and you think of that as a limiting process perhaps that's what they couch it as now the next slide is
00:03:48 what really is uh not obvious and I found intriguing about this it turns out that no matter where you are as a human being if you think of the average human beings in that area we tend to travel in a fairly narrow bandwidth of a a duration we travel about an hour and an hour half a day a day if you're walking to your place where you're going to do work in a farm or in a village in uh sub
00:04:10 Sahara Africa or if you are somebody who lives in the the Bay Area we all seem to travel about an hour and hour and a half a day and uh that's rather remarkable when you consider the previous plot and what we've seen previously about vehicle miles traveled it's almost like this notion of personal Transportation One Way or Another be it by foot or in a in a vehicle or also relates to uh uh
00:04:30 commuting in in buses or other uh vehicles uh seems to be rather constant for Society for individuals um now coming back specifically to Automotive on this chart um I'm going to highlight that in just a moment but this is a layout that was put together by uh Folks at Lawrence Livermore National Laboratory in about 2005 again and they were looking at all of the energy
00:04:50 sources on the leftand side of this plot uh Hydro at the top going down to fossil fuels with oil at the very bottom and the outputs are on the far right are the uses of the energy so what you see for automotive that becomes uh problematic is that the largest component of the fossil fuels that is petroleum or oil uh is utilized for the most part by transportation and that obviously is
00:05:13 something that raises questions in terms of sustainability when we talk about growth in developing markets and and the projected increase in vehicle park car park uh globally in the future so to summarize uh this portion what is I think clear to everybody is you have to have three pillars to uh energy if you're going to have any sort of solution secure clean and affordable
00:05:33 energy is the way I've couched the the terms that are listed here and that's true for every nation states nation states seeking sustainability and what is key I think to understand although there are potential solutions that we all recognize more work is going to need to be done I would classify that as the ultimate solution will come out of advancements in science and technology
00:05:52 and it's why it's an an important area we see more effort going to this this topic so going to now uh uh the automotive industry and in terms of what we're doing today different energy resources again are listed on the leftand side and uh these start out with conventional source of oil if you will to fund our vehic or drive our vehicles and then as you come down to some of the
00:06:16 other more unusual sources and the conversion Technologies and you can argue about this part of the the plot but the conversion Technologies ENT are shown in the center uh Center left and then as you come across to the right you can see the energy carriers that result uh liquid fuels electricity and hydrogen and it tends to be cleaner as you go down to the bottom uh you might argue
00:06:37 that about some of the issues around uh nuclear uh but in terms of on vehicle what you get is certainly cleaner um and then the uh battery technologies in particular that I'll be focusing on tend to influence the hybrids the plug-in hybrids and the electric vehicles shown here um so what's very clear about this kind of chart is that if you're worried about that energy security one thing you
00:06:59 might do is just as you would with any investment you might try to diversify and so you can see that as you Electrify Vehicles you actually diversify the energy streams that are available to you um bringing in for example Hydro or other me methods to get electricity into your vehicle for energy so that's reflected on this chart and it's not it's it's a GM chart but it's something
00:07:17 that other auto companies tend to use as well um and that is if you think about uh what vehicles we have and what's what we're going to see in the future the ordinate reflects improved uh vehicle fuel economy and also reduced emissions uh so up is good in that respect um and then the abissa reflects time so near-term today we're dealing with dominantly internal combustion engine
00:07:39 Vehicles we're coming into hyro electric vehicles they we're seeing a lot of them on the road but nonetheless they're about for example the United States if we're talking about 12 million units a year that are sold there's about 200,000 hyd electric vehicles sold annually today uh in the in the United States uh and then we're moving up into the third generation if you will of uh extended
00:07:59 range electric vehicles that I'm going to talk a bit more about today and hydrogen fuel cell vehicles and reflected with time then as we see these different Technologies come on is the fact that you get this energy diversity and that goes with the last chart that adds security as well as a cleaner uh utilization of energy a very large question and I'm not
00:08:20 going to dwell I have two slides or so on this uh I'm not going to dwell on this a lot for me I I I tend to work in the area of Automotive as an oem so General Motors um we worry about on vehicle emissions now obviously the the total picture is greater than that you have to do a well to Wheels analysis if you will or life cycle analysis so a big question for us
00:08:41 is where do you get the energy from and that's highlighted by the box I placed in the lower part of this chart this was a uh APS American physical Society Workshop that I took this uh uh content for this slide out of the uh agenda for the Sunday session is listed to the left and George crab was Crabtree was the organizer of the session from argon National Laboratory um you can see some
00:09:01 of the talks that were given there um and I took out of George's talk the notion that at the end of the day it would seem like sunlight's going to play a major role in terms of trying to get direct Photon capture if you will or photable take energy sources because they're just so much energy there and if you walk around the slide I won't read them out to you but you know you can
00:09:19 basically see things like at the the the bottom here that the annual production human production of energy is about 1 hour of sunlight don't seem to yeah there we go um and and some of the other numbers around so there's an enormous amount obviously of incident solar radiation that potentially can be used there are troubles and and and uh issues with trying to implement that technology
00:09:37 but it seems to offer at the end of the day a great deal of promise for this purpose so if you look at the bottom when we talk about things like electric vehicles or extended range electric vehicles that I'll talk about today uh there's an arrow that's drawn here on vehicle to the right of this Arrow you have very clean Technologies and high efficiency on the vehicle the
00:09:57 unfortunate thing I'm coming back to again is that when you look to the left here where do you get that electricity from that's the problematic part and that's what we have to work on as we go forward uh so it's listed here in terms of coal gas and nuclear uh fision there are other techniques obviously to get the energy as well um but you want to do no harm going back to the the actual
00:10:15 Source stream as well as on vehicle utilization is uh same sort of argument comes into play when we talk about fuel cell electric vehicles again highly efficient vehicles and the only emission of a hydrogen fed proton exchange membrane fuel cell for example uh is is water and that's a great thing and it does no harm if you will uh but before that time you
00:10:35 got to get the hydrogen from somewhere and as it mentioned here about 90% of hydrogen today comes from the reforming reaction shown from natural gas a lot of natural gas in the United States you might think of it as a bridging strategy uh in in in time we hopefully we'll get to some more efficient way to generate hydrogen for uh Vehicles if we take that route uh one thing that I think is
00:10:55 helpful to just touch on before we get more into some of the technical parks of this talk is is the comparison of gasoline versus batteries uh the the end result is batteries don't store a lot of energy per unit mass or energy per unit volume but it's not as bad as is often made out and that's why I put this slide together so if you look at a battery based EV and this is just real simple
00:11:14 spreadsheet type calculations uh at a battery pack level you can get about 130 W hours per kilogram W hours being your unit of energy uh per unit mass then kilograms and pack uh pack efficiencies over a drive cycle like a federal test procedure drive cycle are about 95% the power inverter about 95% the electric motors that would be used then you're sending your AC current to typically in
00:11:35 today's configurations is about uh 90% 95% it can hover between there so your overall traction efficiency if you will from the battery outs about 80% and you might say then your effective energy density on the vehicle is about 100 W hours per kilogram if you look at Gasoline where you have a great deal of energy it's very hard to compete with liquid fuels in terms of energy per unit
00:11:56 mass or energy per unit volume so about 12,000 W hours per kilogram well that's about 2 of magnitude greater than what batteries provide for you today and it seems like an impossible achievement to try to provide the same functionality on an electrified vehicle that you can get with a gasoline uh fed vehicle but the issue there is that in terms of overall energy efficiencies it's much lower in a
00:12:18 internal combustion engine systems uh because primarily of the heat engine being quite inefficient and so the ratio is about 15 to1 it still points out the challenge in terms of energy density of batteries versus gasoline but it's an utom manage reduced almost relative to the direct comparison of the uh battery itself to the the fuel so to come back once more then to
00:12:40 General Motors and again I think it's reflective of what other Automotive companies are doing as well I've listed our announced vehicle product programs on this chart um it actually starts here way back uh with uh the ev1 I think it's listed up here and um and then our strongest expression of electric vehicles uh the the volt which I'm going to talk more about but our uh our
00:13:03 front-wheel drive two mode hybrids our rear wheeel drive two mode hybrids what we referred to then as our GM hybrid so these are midsize Vehicles high voltage systems and then going back originally to our bus system and again you can see that come about 2010 for all of our programs we moving into lithium ion at that point in time and I think you'll see this again in other Automotive
00:13:21 companies as well um before I dive deeper into the batteries and the Battery Technology and get a little more specific on on some of the technical parts of today's talk um it might be helpful to have one slide that says what are the main challenges for hybrid electric vehicles today and that's listed here and it's what you might expect for those of you who have
00:13:40 uh obviously worked a bit in this area um batteries and Power Electronics and electric machines electric motors are expensive and one of the problems you have when you make a hybrid electric vehicle is you add a redundant or semi- redundant powertrain to your system you already had an internal combustion engine path to a transmission and now you're adding a battery to uh Power
00:13:58 Electronics to Electric machine so you're adding that extra equipment onto a vehicle and the end result is you have to worry about cost you have to worry about packaging that is can you fit all these things in and you have to worry about the mass of the systems there is a mass penalty obviously associated with energy consumption on a vehicle uh then SE secondly you get into issues that are
00:14:16 more engineering related and that's thermal and electrical integration into the vehicle system and last which is quite important especially when we talk about say lithium ion systems is Diagnostics and prognostics for all of our systems on a vehicle today you want to know about the state of health you want to know how well something is doing so that if you need to tell a customer
00:14:33 by lighting a Telltale or uh telling somebody in a service environment what's going on you need to have very robust uh State estimators running to interrogate your various subsystems on a vehicle so those have to be developed for lithiumion systems it's a bit more of an engineering challenge but it turns out it always is domain based and so I'm going to talk a little bit about state
00:14:51 estimators for Battery Systems but to interrogate a battery and determine how much more life it has and whether it's can fulfill its original Mission or not you have to weave in the electrochemistry that governs the system to control models you can't separate the two and so there's a bit of science that still comes into that uh engineering Endeavor um so I'm going to transition
00:15:12 now to talk about the strongest expression of uh you might almost think of it as a hybrid electric vehicle but we tend to use the term an extended range EV for the Chevy Volt the reason for that is uh it has we we call it an electric vehicle and it's it it satisfies certain regulations for example in California California our resources board is an
00:15:31 electric vehicle because no matter what you do with the vehicle which is coming out in December of 2010 no matter what you do uh for the first 40 miles after a full charge it will run as an electric vehicle you can put your foot all the way to the floor and the heat engine will not come on now in a conventional plugin hybrid electric vehicle what happens there is if you pull onto an
00:15:48 on-ramp for example and you put your foot to the floor for the accelerator pedal the heat engine will come on in addition to uh Power being delivered through the battery pack to the Power Electronics electric machine so you have both and both systems operative and you won't have an electric vehicle at that time because you are running your heat engine on a hard XEL so in in this case
00:16:06 you drive off your 40 Mi and after that point in time the rest of your range is uh treated as you would for a charge sustaining hybrid electric vehicle where you have both your uh battery working and your internal combustion heat engine working and the batter is in a charge sustaining mode at that point in time um the intention here is to overcome the original range anxiety of
00:16:26 the ev1 shown on the left so an electric vehicle but couple that with the fact that you truly do get for the majority of your drives people who buy this vehicle 40 Mi of true zero emission vehicle range um so this chart tries to exemplify or indicate why that's a value and the demographics point out that 80% of all customers 78% commute 40 Mi or less
00:16:49 daily uh now obviously the people who would buy this vehicle would uh be those that are on the high end of uh short vehicle ranges and therefore they may actually never even go to um a a fill up at a gas station unless they take rare trips that are quite long in which case the heat engine would offer them the extended range after they drove off their electric
00:17:10 range um so I thought what might be helpful rather than try to describe this thing is just show a video and um hopefully this hopefully it's going to going to run actually here okay battery pack so it's bu you can build around the battery pack rear of the vehicle coming on front of the vehicle coming on that's the heat power Electronics then the heat engine
00:17:30 um you can see things being flushed out then the the chassis parts are coming on close out body pan uh coming again back to the rear wheel the rear wheel close out uh that's the dash coming on and seats and what have you um I'm going to show it once more and then I'll make a couple comments on it so the main thing is that you you pretty much have to start these
00:17:54 things I think I'm going to show Once More you have to start building around the main structure which is the battery pack by mass and um that's that's basically the the gist of the video and then you can see that the everything else is basically publ uh packaged in conventional volume space of of a vehicle and it becomes pretty standard
00:18:19 after that um you know some of the things you want to keep in mind is batteries are a lot like people you want to give them the same temperature about you want to treat them the same in terms of Crash environment you want to put them between the frame of the vehicle you want in any sort of uh uh uh I'll call it limited event crash the battery pack doesn't get crushed um and and that
00:18:38 then helps a lot in terms of the uh uh Energy Management on the vehicle and in terms of uh preservation of structures and people okay uh so uh one other thing to keep in mind we talked about the whole weld of wheels and the question in this particular case from an eev extended range electric vehicle point of view is what impact would we expect that to see
00:18:57 to have on the utilities and fact of matters it' be very little now this is I I should point out this is a bit of a flip it's too strong word but it's it's sweep some things under the rug um 10 million Vehicles is a lot in and in fact because of the depressed annual vehicle sales volumes right now um that's actually nearing about what we're selling these days but uh the fact is as
00:19:17 showed on the previous chart that's not the total vehicle car park so you'd over time then start replacing the total vehicle car park the point is that in the near term you wouldn't see an impact on the utility grid and that's the intention of of showing in this slide that's less than 1% of the utility grid consumption right now uh that you'd have if you if you put 10 million eevs on the
00:19:36 veh on the road um so what what's driving lithium ion and um uh what I just want to point out is that I I think it's clear to folks who have worked in The Battery area but it may not be as clear uh to everyone if you look at the turn of the century before it was pretty much Le acid was the dominant battery that was rechargeable and the Workhorse of of uh uh certainly the traction that
00:19:56 is a vehicle use of battery uh industry and then what happened in about 1990 the inventions that allowed for this the science came out earlier uh you well I'll Show a slide on that but um insertion electrodes came out and metal nickel metal hydride batteries are insertion electrode there you're inserting hydrogen in nickel oxyhydroxide nickel hydroxide electrode
00:20:16 or on the other side of the cell the U uh metal hydride system and then after that shortly after that lithium came out mainstream again an insertion electrode based system insertion electrodes typically have very high cycle life you actually are controlled by solid state diffusion as far as your energy transport goes or you can be at least uh especially for larger particles and uh
00:20:38 they're very different than the film forming electrodes of lead acid and other uh similar electrodes of the past um uh materials reseearch has played a dominant role in bringing these uh systems uh into play you see the uh graphine planes stacked up to make the uh typical negative electrodes shown there and the metal oxide and metal phosphates then are indicated as well
00:20:56 for the positive electrodes so before before I go too far I I realize the audience is somewhat diverse so I'm going to dive into some stuff that's more uh specific to electrochemical research and and that's not maybe going to keep everybody but before I get there I'm hoping I can keep some folks uh uh by saying here's how a lithium battery works um my son is is 18 and he made
00:21:17 this slide up and so I I'm hoping that it works well here as well it's it's animation and it gives schematically an idea as to what happens in a lithium battery remember I said insertion electrodes so what you have is lithium is inserted in one side and it's pulled out goes across an electrolyte phase which is the yellow phase here and it jumps from SES in the electrolyte phase
00:21:36 that are salt sites if you will and moves across uh to give rise to uh uh ionic communication through the electrolyte phase and the electron goes on the outside circuit and does work for you through electric machine and so that's what's happening right here you can see this is on charge actually and if you went on uh discharge you just see the exact opposite the electron would go
00:21:57 the other way and the lithium ion would go the other way um and the reaction that we typically use then is shown uh up at the top in the red box and so you see that lithium ions combined with an electron uh on charge and and that's happening at this electrode over here and then a vacant sight which these would be vacant sights inside of the uh uh host material and that gives rise to
00:22:19 this intercalation species here the lithium ions have a slight positive charge and then the surrounding carbon atoms in this case have a slight negative charge and that's indicated by this uh Kai so now if we take that same plot and we try to expand it to see what's going on we'll go we'll try to dive down into the atomic level if you will as to what's
00:22:36 what's happening it's quite analogous at each electrode so right here I'm looking at one electrode and the total cell width is on the order of uh 100 to 200 microns each electrode is on the order of 50 to 100 microns depending on whether it's a high power electrode or high energy electrode and the the electrodes themselves are actually uh composed of uh small particles they
00:22:56 could be spherical or they could be platelets or rods or other shapes and uh that's uh shown over here on the right lower right and then if you look at one of these particles and expand what happens inside there is this is a negative electroc carbon host you see these graphine planes you would see metal oxide planes for conventional positive electrodes or spinel
00:23:17 threedimensional networks but the lithium can travel inside of the uh inter interstitial sites or inside of the uh layered regions and insert and therefore be soaked up like a sponge and and uh give rise then to this insertion electrode concept so um I'm going to make one comment here this goes back to some of my own work and and others have
00:23:38 obviously worked a whole lot in this area make strong contributions um but one of the things that happened in the uh early 1990s as many of us were looking at lithium metal uh negative electrodes instead of uh intercalation electrodes and Sony was the first to come out commercialized and intercalation negative electrode based on carbon um actually and I have this at
00:23:57 the bottom of the slide if you go back and into the literature was back in the 1800s uh when people were looking at intercalation of graphites with various species anions and cats and so what happened in 1990 or so as people realized well you could use these materials to actually make a negative electrode in a lithiumion battery what also greatly uh accelerated the use of
00:24:16 uh uh lithium batteries in various applications was the development of a stable non- aquous solvent it's a product doesn't have a proton like in water that can react and uh that was done actually at at at Berkeley uh uh 1959 was 58 was a thesis um I mentioned this to some perhaps it's helpful to some folks that are wondering about Industrial Research and how things are
00:24:36 done there um a lot of questions have to be answered I would argue rather fundamentally to allow you to decide what path to go on an applied path so this uh work was actually done in the early 1990s by myself and my colleague uh Brian Cook and what we did is we took one of these polyol nital fibers shown on the left and we could easily make a micro electrode out of it just basically
00:24:58 making a attachment it's not hard to do with these six Micron uh diameter fibers and then we could cycle that individual fiber and the current response to a cyclic potential Source this potential source is shown in the abissa here is shown and this is for 300 Cycles taken digitally then and just plotted over one another and the point is it's extremely stable and it's traveling the carbon
00:25:17 fiber and lithium lithiation process over the potential of interest for lithiumion battery and what it allowed us to say to our management is we're having all kinds of problems with lithium metal negatives everybody else was as well and uh these are very stable systems and we can actually make batteries out of them that will work and so we switched
00:25:34 our own program internally largely based on this kind of work and so it it allows you to do scientific work if you will that is unambiguous in its interpretation and can drive then robust decision-making so uh what's happening now in lithiumion batteries today well the conventional system is the 4volt system indicated here so i' I've lay I've listed a layered metal oxide versus
00:25:55 a carbon negative and gives about 4 volts and then more recently people have looked at trying to improve the stability of these systems mostly for life and and that means using a lower voltage positive and a or a higher voltage negative titanates and metal phosphates and so that narrows it actually reduces the cell voltage you can then you then reduce unfortunately
00:26:14 the cell energy voltage of energy being proportional to the cell voltage but you get very high stability and that's needed for applications um and I'm going to come back to that point in just a minute um in eevs versus plug and I mentioned the Chevy Volt is an eev an extended range electric vehicle what you have as a major difference then versus plug and
00:26:35 hybrid electric vehicles is shown here you just have to have much more power if you're going to have it so that you have full function electric vehicle when you step on the accelerator pedal uh at all times for that first 40 miles that means your electric machines and your battery pack have to be scaled up to maximum power so about 100 Kow for a very relatively small vehicle which the volt
00:26:52 is relatively speaking and and that's the main difference in those systems a very high power to energy ratio capability relative to for example United States Advanced Battery Consortium requirements and goals for plugin hybrid electric vehicles and the Big Challenge is then our life and cost for these systems um now I showed you that people are going
00:27:10 into metal phosphates for durability and I mentioned the last slide that life is an issue so one of the things we would like to do is start looking at whether these phosphate based systems would actually work well in extended range EV systems and I'm going to transition to that part of the talk now one of the problems you have and I mentioned earlier is State estimation you have to
00:27:27 be able to tell how the system is behaving and this plot makes clear how easy it is to tell that usually what you do then is you read off the voltage and these are all taken at very low scan rates all of these voltage traces on the right so once you read off the voltage you can follow across let's look at a super capacitor the red line if you read off a voltage and you come over here and
00:27:46 you can see that you're at about excuse me about uh 60% state of charge when I read off a voltage of about uh uh 1.6 volts and and so that's a thermodynamic lookup and it's actually a function of temperature as well but you can correct for that in your lookup tables and your vehicle controllers it turns out those two systems that I mentioned that are of Interest today for their durability uh
00:28:08 the tight Nate system shown by the slight purple curve and the orange curve slightly higher voltage metal phosphate system have a very uh horizontal uh voltage profile and that proves to be problematic so on this slide it gives you a really brief tutorial on how you construct a state estimator and why that shallow slope if you will is is is problematic first thing you do at any
00:28:29 point in time is initialize the state of charge you have to have it either from a lookup table because you keyed off on an event and you had a state of charge recorded then you start from there or you have to infert by an open circuit voltage at startup but you initialize your state of charge in a vehicle controller then as soon as you start driving or passing any current you
00:28:46 record your current your voltage then you can do two things you can first do coolum counting that's pretty easy just count the number of kums or passed through on your current sensor and that's how much charge was transferred from one Electro to the other and you get the state of charge of your system how much energy you can deliver that's not adaptive however that's just
00:29:01 counting and current efficien are never 100% so you can't do that forever it gets off and it never adapts back to getting corrected so the Adaptive part is the voltage based and that's shown here what you do is uh you use a very simple model typically lumped parameter model that retains the Salient features of the electrc chemistry you regress parameters adaptively for that voltage
00:29:22 based model and that gives you the state of the system and once you've gotten that volt based model working one of the parameters you regress is the open circuit voltage well that then goes back to your lookup table shown here on the right and again once you have that you can get an S SOC so now you have two soc's one is a voltage based state of charge measurement one is a uh kulum
00:29:42 counting based state of charge measurement and then you have a blending algorithm people in sensing uh sensing uh algorithms I'm hope I don't offend anybody here but when you have something that's ad hoc you usually give it a fancy name so it's called Sensor Fusion uh you you you fuse these different uh things you might use a weighted average you might try to have some sort of
00:29:58 conceptual way to do it there are ways that are a little bit better than others and and some of the uh Hallmark algorithms are based on the work of Colman and Colman filters but at any rate uh you can see from this chart that this voltage based part which is your only adaptive part is problematic when this slope is near zero the open circuit voltage here versus state of charge so
00:30:16 what I'm going to talk about next is a way to deal with that that's the motivation for the next part of the talk so what we did was a chemical modification this is from some work with uh uh hrl colleagues that had mentioned earlier and and we have a a patent pending on the top of this that I've listed there in a publication that's been accepted for publication actually
00:30:34 by Journal El Chemical Society that's listed um but in essence the top portion here is your conventional metal phosphate system so you have a metal phosphate positive electrode shown here lithium iron phosphate and then you have your graphite and you always have an excess of your graphite because you don't want to Plate lithium on overcharge shown up by this reaction and
00:30:54 actually on over discharge what can happen is there's a copper current collector associated with your graphite electrode and if you over discharge the copper dissolves that's not a problem actually when it dissolves the problem is when you go subsequently to charge your battery that copper plates out it forms dendrites and you short your cell and that can get exciting so uh what
00:31:12 what you want to do what we did then was to say well we'll put lithium titanate which is a very stable material and it will have a different voltage signature and we'll put it so it comes in just at the end of discharge so it's a state of charge marker towards the end of discharge and allows you to make sure you have enough energy left in there so you don't do something naive like come
00:31:29 up to a stoplight in a vehicle an eev and say do I turn off or excuse me when you're running in the charge sustaining mode as an example do I turn off my heat engine or not my air conditioning system is running it drains at about 1 kilowatt if you turn it off and it turns out you're near the end of your state of charge you can end up falling off not being able to have enough energy then to
00:31:48 start your car afterwards that's why you need to know uh the state of the system for efficient vehicle operation also for energy consumption calculations that are done real time on vehicles so uh I I should mention one thing I when I was uh putting together this talk I was also trying to keep up in the literature I happen to notice that John and colleagues uh uh here uh Paul albertis
00:32:08 and and Jay Christensen now uh now at Bosch um we're looking at a similar problem that was uh they were looking at multiple uh constituents in the positive electrode I'm talking here about the negative electrode but they had some similar similarities I mentioned here in a quote that you can actually uh get better handle on the state of charge or state estimation if you will of uh
00:32:29 systems by mixing different materials that give rise to a different voltage response um so coming back then to the negative electrode that that I'm talking about the titanate addition to the graphite negative electrode um one of the things we had to worry about was the electrode stays at one potential so you have graphite and you have titanate on the same electrode but both of those
00:32:50 species see the same potential normally titanate systems don't go to very negative potentials near that of lithium you typically operate them at about 1. 5 volts positive to lithium that's one of their strengths is you don't drive down towards the lithium potential but if you're going to utilize all of the capacity of the graphite now your Titan has to survive at this very negative
00:33:07 potential very reducing environment so we want to know if it was stable and the the end result of this plot with the X-ray data and the cych life data convinced us that yes indeed it does seem to be very stable at these very negative potentials the the tight Nate itself independent of the graphite and uh the I'm sort of jump into some of the conclusions here because I think it's
00:33:25 kind of clear enough that I don't probably have to go into some of the uh uh more detail for this talk but this is the the top plot here is your conventional it's an A123 cell actually so your conventional uh lithium iron phosphate graphite cell and you can see again this issue that you get an extremely horizontal voltage Plateau at a fairly low state of charge now this
00:33:46 isn't that low it's C over5 so you discharge the battery in this case uh at uh 12 minutes for total charge and total discharge at this current so it's it's fairly High rate actually but it's it's slow enough for this system it looks like almost thermodynamics in terms of no hysteresis or little hysteresis now if you come down to the second chart here on the bottom what you see is that
00:34:05 as you are discharging the cell instead of having to just drop off and then go back it drops down to the lithium titanate potential hangs around there for a while so this is the step you have it's a very clean step and you can tell with thec marker what's going on and then you reverse the whole process and come up on the top upper curve and and recharge the battery so you have a clean
00:34:24 way to estimate the state of the system how much energy is in there and there are other ways to inter arate uh the effect of the lithium titanate that I'm showing here um and and this is just cyc voltametry um and and you can see that the bulk of the capacity is carried over here by the graphite uh iron phosphate couple but a little bit of capacity is carried over here by the lithium
00:34:43 titanate system as well um so one way in order to figure out if everything's working as we'd expect I mentioned that there's one potential and you've got two electrodes electrod materials so uh this is a bit uh uh edisonian as far as how we're doing it but allows you to step through and logically see what's going on we first just put the two Electro materials
00:35:05 on the same electrode uh both connected to the current collector so the same voltage and uh sure enough what you find out is you can just use a very simple relationship that would come straight forward uh come right out of thermodynamics for the carbon and the lithium titanate and if you add up then the individual curves you should get that's called the model here but you
00:35:26 should get what you get in the experiment sure enough you do so it's behaving as an ideal system there's not any unusual interaction between the lithium tightening and the graphite system you can sum potentials basically once you recognize how much charge is going to each uh constituent and now this one is different than the last one now we mixed the powders and made up the
00:35:43 electrod so they're intimately mixed and indeed you get the same sort of thing if you know how much of each powder is there um you can go ahead and construct a very simple model and you can represent uh the the results that are in the red with the uh simple model uh it's it's too simple even really called model theramic relation that you expect in the uh uh blue
00:36:02 curve okay I want to touch a little bit on some promising developments and then I'm going to come back to what I talked about earlier on stress strain and degradation and um I realized I'm going to be challenged on time I should be done by noon I think is the idea or quarter two for questions noon noon and then 15 minutes of questions and then people get to go
00:36:23 eat at that okay good um all right so this is why uh you know I I I my thought is I'm very optimistic on where batteries are going to go they're they're likely going to have a profound increase eventually in energy density and it's exemplified by this article others have uh published similar uh pieces of work this is from the group out of Baux cnrs Laboratories
00:36:44 and the capacity of most systems is most systems of interest for the negative electrode are plotted here Milli hours per gram so that translates to it's proportional to koloms or or charge per unit mass of material and you can that we're using today uh uh lith carbon about 372 mlion per gam that's the theoretical value Li I6 and then as you move to the right you get much higher
00:37:07 capacity materials that we aren't using today people have tried them and the problem is as indicated by this large arrows these things tend to decrepitate or break apart because they have very large volume changes uh silicon right now appears to be the the the clear single U uh Matrix uh winner if you only use one material now that looked problematic uh shown here uh or it could
00:37:27 be be problematic in terms of the expansion it's about 400% or so at full lithiation of the fully lithiated silicon alloy however if you're not greedy so to speak and you don't use all that and you combine it with uh carbon particles it turns out you can get very high robust cycling very high capacity robust cycing that's shown here about 1,500 milahs per gam remember that's uh
00:37:49 as compared to today's in a conventional lithi carbon electrode 300 Mah hours per gam usable today so there's going to be a lot of progress there the way people are doing this as as mentioned here is by going to nanop particles and you can think about it because it makes sense when you think about it solids are are solid because they like themselves otherwise they
00:38:08 would they just float apart they wouldn't be a solid and what happens when you go to these nanop particles the adherence of atoms to themselves at the surface tends to become dominant because you actually get when when you get below about 10 nmet or so you've got more atoms associate with the surface than you do in the bulk and so the materials tend to hold together and so what is
00:38:26 characteristic of making high capacity electrodes like this work is going to these Nano structures where the surface tension plays a dominant role in holding the structure together and you can tolerate the expansion and contraction and I'll I'll say a little bit more about that um so this is some work now coming out of the Argon National Laboratory group um through the
00:38:44 uh work with the Department of energy and link to the B program here um but uh what's of interest is that today's positive Electro materials that i' I've talked about metal oxides and uh TI uh uh spanel structure are typically about from a practical point of view 150 Milli hours per gram the materials that the Argon group has come out with and they're being looked at by various
00:39:05 entities uh uh around the world now uh Envia NVA actually out here in Hayward California is looking at it as well are uh getting much higher than that about 200 to 250 milliamp hours per gram and it's an interesting tailored structure that has an inert phase li2 M3 that is of a rock salt structure and is contiguous in fact you can't see the difference in under temm with the
00:39:27 Associated active phase which is a layered metal oxide and it gives very robust Cy it appears to give robust cycling performance and it appears to be something that will be uh quite important for us in the future um another new area that I think could be quite important is uh on the positive electrodes a lot of work has gone on in the metal phosphates another linkage
00:39:46 that is extremely stable and very low cost are are silicates and so people are starting to look at metal silicates that are already up to fairly good cycle life at 100 million power per gram these should be extremely low cost materials uh metal phosphates are great except they do have a fair amount of cost structure associated with them the metal silicates might be quite promising for
00:40:04 driving down that cost structure so um I I want to speak a little bit about life and get into some of the durability issues um there are two problems that give rise to life issues in lithiumion batteries one is chemical degradation and one is mechanical degradation and I'm going to talk first a little bit about the mechan the the the chemical degradation it
00:40:23 turns out that affects counter life if you just let a cell sit there it will chemically degrade especially so if you heat it up if you cycle it this expansion and contraction gives rise to cycle life problems the two working together will unfortunately give rise to combine problems so what I'm talking about is the wear out mechanisms to the far right of What's conventionally
00:40:41 called the bathtub curve uh you have infant mortality at first for failures then you have a long time in service in automotive we typically talk about 10 to 15 years we have a 10-year warranty for example on the battery pack and the volt that's coming out and then you have wear out so uh in in the wear out area first we're going to talk about the chemical degradation what the reason a lithiumion
00:41:02 battery works is the solid electrolyte interface now there are other things that make it work too but without the solid electrolyte interface it wouldn't work so I want to talk a little bit about that this is some very high surface area carbon and it's it's nice to use for this study because it it it it accentuates the effect of the first cycle inefficiency what happens on the
00:41:19 very first cycle shown in this first panel here is when you scan the potential negative to A reduced environment you you break down by reduction your solvent you have solvent reduction you see gas coming off in this case propine gas because it's a propine carbonate solvent and then eventually get over to the lithium reaction here and then afterwards and it's just
00:41:37 amazing how this works it doesn't you don't see any evidence of it anymore after that Cycles 2 three and four are here and they're very reproducible after that first cycle where you've conditioned the surface and no more solvent reduction from a appearance point of view is happening well there are various uh mostly FTI but various uh methods by which you can interrogate
00:41:56 what's on the surface and and and the uh uh dominant species are talked about here and uh there's a uh pretty good agreement that actually this is the the uh dominant uh species that's the organic layer and this is a function of state of charge because if you look at this reduction reaction you can combine these and look at that as your state of charge more lithium in your system
00:42:16 refers to a higher state of charge in this negative electrode um so you're breaking the bonds of the propylene carbonate molecule here and the same thing happens with ethylene carbonate that are in conventional batteries today it turns out there's a layered structure to this SEI the inner layer appears to be a very compact lithium carbonate region that's what's shown here when you
00:42:36 first start making the layer and then the outer layer is this more gelatinous amorphous uh uh lithium adex layer that was shown in the previous slide and this is some work by uh uh tem uh that looks at that uh initial lithium carbonate and now what you can do after you've grow in the thicker layer is you can use AFM in the tapping mode and actually scrape off a little bit of the uh layer and and get
00:42:58 the total thickness and that's this organic species that you see by ftir and so it's about 20 NM uh in thickness at this point in time when you have this organic layer on top these are protective layers now I should mention before I go on that at the positive electrod similar phenomena goes on it's not as well studied it is important it's just something that people haven't
00:43:16 gotten around to but if you go back to some of the early literature what you can see is that these cyclic carbonate ethline carbonate proplan carbonate they tend to be reduced on highly catalytic materials Platinum for example and nickel oxides as well but Platinum it will be reduced at about 2.1 volts versus lithium we need those systems to operate up around 4 Vols versus lithium
00:43:35 and the reason that they operate there is there's a kinetic limitation at the interface because these very reversible materials will actually cause the reaction to occur to lower voltage um so the point I want to make that's not obvious to folks is that if you just went by the thermodynamics that is that solvent reduction reaction on first cycle it can happen it gets poisoned by
00:43:53 subsequent SEI formation and same thing with that positive electrode the other electrode you can get solvent oxidation it can happen it just surface phenomena stops it from happening but if you subtract those two voltages uh the 2.1 minus8 you got a 1.3 volt battery versus the 4vt battery we typically operate with 3.5 to 4vt battery so that underscores the importance of the solid
00:44:14 electrolyte interface the second point is I mentioned these are the same plots I had in the previous slides uh it looks like there's no solvent reduction going on out here uh uh after this first cycle the reality is that there is and it's consuming lithium the whole time each solvent reduction reaction is as is shown here consumes lithium if you think of the first cycle What's Your Capacity
00:44:37 all of your lithium starts out in the positive electrode it goes over as as made and then when you put the cell together it shoved over to the negative electrode sometimes they call these rocking chair batteries then you shove the lithium back over to the positive electrode if you lose lithium as is shown by this degradation reaction you lose capacity you can think of that
00:44:53 initial lithium put in the positive electrode you're losing it you won't have as much fuel to go back and forth eventually you have none you won't have any cell voltage you you lose your energy so what this shows is that as you it's like compound interest if you will as you start losing lithium to end Cycles you can dve a formula and show that for like a volt program our our our
00:45:12 programs com out in December 2010 you need 5,000 cycles and what does that mean in terms of current efficiency what it means is that if your current efficiency isn't above 99994 49 you won't make your 5,000 Cycles you will lose 20 % of your cell capacity and you'll have failure early failure and in our case it means warranty and that affects job security
00:45:34 and other issues for me so so the the issue is you have to understand actually you are getting solvent reduction out here you just don't see it it's so low the current efficiency for your dominant reaction is so high um the the next item is the the uh uh expansion and contraction and I'll try to do that in the the last 10 minutes here um but this is where you combine now this chemical
00:45:55 degradation with expansion and contraction this again my son made this chart up um but it gives this 10% expansion when you put lithium in and in contraction in sort of a graphic form you can see this in graphite this is graphite flake you can see it by Ramen uh that you get cracking you can see it by uh Peak broadening uh x-ray defraction you can see then the
00:46:13 crystallite sizes uh tend toe decrease with cycling and so you do get there is strong evidence for cracking of uh graphitic electrodes this is a from a the group at Toyota where they're looking at the positive electrod Toyota uses lnca and some some of their experimental test fleets that are actually operating in in Japan today and as you put lithium in as is shown by
00:46:33 this lower plot uh it's a little bit convoluted the way it it's plotted here but basically putting lithium in goes down here uh or I'm sorry taking Lithium out so you get a 4% contraction and then an expansion as you put lithium in and and that correlates with what you see in terms of this cracking phenomena that goes on in the uh lmca electrode that is shown lithium nickel Cobalt illuminate
00:46:54 electrode so you see it on both the positive and negative electrod and one analogy that I think is helpful is to think about what we see and we all kind of are familiar with with mud flats or or sand when the rain comes out and it hits the ground and over a mud surface the surface expands that's just like the lithium being shoved into an electrode you're shoving the solute in and the
00:47:16 host Matrix and the solute that's combined with it expand and you don't see any cracks and then as you take the lithium out on discharge it's just like water evaporating when the sun comes out and you get the cracking so it's on discharge events after a fully charged system that you tend to get a lot of surface tension that pulls apart the surface and gives rise to cracking and
00:47:33 that's why I put the uh the picture of the mud flats in there we see it on iron phosphate I won't go through the chart but the lefthand figures show that as you cycle things you get cracking in iron phosphate systems and they expand and contract again about 7% as as shown by this article um this is a turns out another piece of work by by John uh and and uh colleague Christensen and they
00:47:54 were looking at uh some early work in this expansion and contraction area and what I wanted to come back to is this theme that I mentioned before if you're going to go to high capacity electrodes getting small is going to be important so the different stress components you have to consider the radial and tangential stresses are shown here and this potential for fracture is maximized
00:48:12 at the surface when you have a full charge equilibrated full charge and then go to a very strong discharge event and get that that tension over the circumferential forces of the surface and what you find if you look for example at the radial stress same thing with the tangential stress is that as you go to very very small particles surface forces can actually draw that
00:48:31 that uh the the stress distributions to more compressive stress because the material likes itself if you will by the surface energy um so uh I won't go through the theory in depth other than to say you can work out some very simple classical equations and do one perturbation on the system that is put in surface energy and surface modulus and show that indeed as
00:48:51 you get smaller with reasonable parameter values as you get smaller about in the range of below about 20 NM these surface tension effects start to become very important and try to hold the uh and and T tend to hold the particles together um and to perhaps underscore this is a bit of a empirical uh non-scientific uh uh support there was a
00:49:12 uh multi-day Symposium at the recent electrochemical society meeting where it was just on nanostructured uh materials for energy storage and conversion the basic notion that you saw time and time again there was that people were getting higher cycle life when they use very small Electro geometries um so the total picture then for degradation is I'm trying to tie
00:49:31 together now the chemical and the mechanical degradation in these batteries which is what again drives us towards things like warranty and and understanding life of these uh systems is expansion and contraction giving rise to cracks at the surface passivation of those cracks any new material on the negative electrode has to be passivated forms SE otherwise you don't get a
00:49:49 working battery so it's passivated that's different passivation reactions are shown here uh and then in those those passivation reactions you have loss of lithium that's part of the passivation reaction that's loss of capacity you can also add in then the omic drop that comes at the surface the irreversible drop from that surface layer that's covering the
00:50:08 electrode the active SES and you can even get to more profound problems of isolation of active materials shown on the lower left so one thing I just say in terms of the life issues uh counter life and cycle life calendar life would favor going to very large particles that minimizes the surface area for this chemical degradation however as I just talked about if you
00:50:29 don't have nanop particles or very very small particles you can be hurt on expansion and contraction Pyle life so cycle life I would argue tends to favor very small particles and so philosophically you might say life is a balancing act between the two um and you obviously want to reduce both of those um I'm going to uh draw the summary here because I think otherwise
00:50:52 I'm going to run out of time um and mention to you the you know recap what we talked about today so early on uh some issues on energy sustainability and some of the problems that we face as a society and and some of the potential Solutions will likely come out of energy diversity at least as a bridging strategy but the importance still of identifying where our initial uh
00:51:11 material is going to come from for uh purposes of of energy consumption uh then we looked a bit at electrochemical energy storage Technologies focused on lithium batteries and extended range electric vehicle applications um I want to thank again my uh colleagues from HL Laboratories and YT Chang that I've been working with at University of Kentucky um and last we talked a little bit about
00:51:32 some of the promising developments for Next Generation liim mind batteries so I'd be happy to answer any questions now thank [Applause] you why doesn't GM make the e one again well you know in a sense they are I mean the people who worked
00:51:59 like myself on the ev1 are the same people do in the vault um so you might be asking why don't they make a pure electric vehicle without an internal combustion engine and it I'm not a marketing person the general thought was this range anxiety was too difficult to overcome for a marketing especially in cold temperatures you might do it in San Diego or you know other places but if
00:52:17 you're going to sell throughout the nation that might be problematic so the notion here is let's try to do an electric vehicle 40 Mi will satisfy my drives most people's drives um if if they want it to be an electric vehicle they're going to get one if they want to also just take off at some point for uh you know uh longer trips they'll get that as well so hopefully will satisfy
00:52:36 uh more people so there's a lot of discussion about usage of a used recently to do some buiness which is not good for automotive anymore but can be used for the other uses so give the technology available today can you comment on like how many usage will be appropriate for car and what would be the yeah it's a it's a
00:53:05 really big question dual use we want to see that too we've we've worked tried to work uh towards that in the past with uh the oems with Epi uh or other entities that uh might use the batteries uh the idea is that you say you have a Chevy Volt these battery packs are very expensive uh you lose 25% of Your Capacity end of life well the battery pack still has 75% of its capacity for
00:53:26 many sufficient for application so sell it then secondly at a much reduced price but buy down the initial cost of the battery by having this dual use one of the issues there that is of primary concern right now is that from the OEM perspective I'm going to sound like I'm whining here but uh part of our issue is we can't get good requirements from utilities or other folks to say okay
00:53:46 here's what I have to have because then we could say okay uh at this point in time swap out the battery for the customer put in a new battery sell it to the utility because it meets their requirements still cuz State estimator is characterizing in the state of the battery um so my sense is this is a a a short-term issue in the long run I think we will get dual use it makes sense it
00:54:07 uh if it's going to work here it's still going to be have some some life afterwards so it's it will be important coming down the road thank you very much to present you with this poster to remember us by let's thank again
