bulletin cover story A123 Systems: Eight hard years for an ‘overnight success’ An interview with Yet-Ming Chiang

A123 Systems is currently one of the leading advanced-technology bat- tery companies in the United States. The company, winner of the ACerS 2009 Corporate Technical Achievement Award, manufactures powerful lithium batteries based on a proprietary nanophosphate technology. Although A123 has been in existence only eight years, its products already have made inroads in the power tools, transportation and energy grid sectors. Its batteries can range in size from small packs for rechargeable drills to trunk-sized kits for hybrid automobiles to tractor- trailer-scale systems for electrical utilities. Although privately owned since it was cofounded in 2001 by ACerS Fellow Yet-Ming Chiang along with Bart Riley and Ric Fulop, A123 made business headlines just weeks ago when shares of the company were first offered for sale in the stock markets and A123 Systems cofounders, from left, Yet-Ming Chiang, Bart in an eagerly anticipated initial public offerings. Riley and Ric Fulop, along with CEO David Vieau. But the IPO wasn’t the only big news for A123 in 2009. In August, it was one of only a few U.S. companies selected Chiang, who is a full-time professor at MIT, was recently in- by the Department of Energy to receive a major grant to help terviewed by Bulletin editor Peter Wray about how he became them evolve into businesses that can compete worldwide in the a ceramist, the evolution of the science behind A123 batter- vehicle advanced-battery market. ies and what it’s like to be both a researcher and successful entrepreneur. ACerS 2009 Corporate Technical Achievement Award an you explain how you first The Society established its Corporate Technical Achievement Cbecame interested in materials Award to recognize a single outstanding achievement made by an ACerS corporate member in the field of ceramics. and ceramics? Indeed, A123 Systems’ nanophosphate technology has I came to the U.S. when I was six. shown significant merit and represents a gain to society through its commercialization. My family settled on the East Coast This year, Yet-Ming Chiang, representing A123 Systems – in New Jersey. I ended up at MIT winner of the 2009 CTAA – will make a presentation on the in 1976. My sister was already at award-winning technology at 8:20 a.m. on Tuesday, Oct. 27, as part of ACerS Emerging Opportunities for Ceramic Sci- MIT. I decided to major in materi- ence and Engineering Session. als, because, even in high school, I Chiang will also participate in a panel discussion during the Opening Session of MS&T’09 on Monday, Oct. 26. The panel had an interest in metallurgy. I did will begin following the keynote presentations. a lot experiments in my bedroom Both events will take place in the David L. Lawrence Conven- at home, including, for example, tion Center in Pittsburgh, Pa. learning how to case-harden steel

American Ceramic Society Bulletin, Vol. 88, No. 9 23 Yet-Ming Chiang Interview parts. I learned about hardening, received tenure at MIT, I started look- develop, how to because I liked to make machinery and ing around for things to work on. Most design and how little tools and gadgets at home, and of my work up until then had been on to engineer these I found that they actually wore out. I the fundamental aspects of ceramics, materials for bet- found this concept of hardening and but I started looking more for research ter batteries. thought, “What a neat thing.” opportunities in areas the DOE today What was the I found out only after coming to would call “use-inspired” research. I time frame this MIT that what was actually going on started thinking about use-inspired was going on? can be found in the phase diagram, that materials. Materials particularly for I believe that allowed this to happen. I learned about new technologies. In the mid-1990s, I our first papers the iron–carbon phase diagram from started to become interested in lithium published on professor Robert Olgilvie. As I majored battery material, because there was an battery materi- in materials science and engineering, a eye-opening collaboration I had at that als were in the lot of the really interesting research at time with three other MIT faculty, mid-1990s, in the that time was in ceramic materials. Gerbrand Ceder, Donald Sadoway and 1995–1997 time I started working with David Kingery Anne Mayes. frame. My work as an undergraduate. I was a junior at What I started to learn was that with A123 started the time, but I spent my sophomore many of the materials that stored in the 1999–2000 summer in an internship at a West lithium were, in fact, oxide compounds. time frame. In Coast company. They really didn’t These were ceramic compounds. They my lab, we were know what to do with MIT undergradu- weren’t much thought of as ceramics, working on a ates who were relatively untrained and but they were used in an engineered- couple of different A123’s aftermarket Hymotion packs can convert standard hybrid automobiles into plug-in vehicles. unskilled. So I spent much of the sum- powder form. The fact is that lithium concepts related Some plug-in hybrids can achieve over 100 miles per gallon of gasoline mer enjoying the California sunshine cobalt oxide, then the baseline material to batteries. One but also coming to the realization that for positive electrodes for lithium-ion developed as I learned something about I’d better go figure out what to do with batteries, is an ordered rocksalt com- how batteries were made. It’s a fairly my life. When I came back to MIT pound. What was interesting was that, straightforward process. that fall, I decided to try out a couple having done earlier research in my Let me back up here. First of all, the of research projects. One of them was career on magnesium oxide, a rocksalt vast majority of the materials used in with David Kingery, and, in pretty compound, I looked for interesting batteries today are powder-based. These short order, it really developed into a technological examples of rocksalt are powders, maybe ceramic powders, passion of mine. compounds. Frankly, there weren’t that carbon powder, occasionally silicon or I was supervised at the time by the many. Spinels, for sure. Perovskites, metal powders. There is a lot of powder late Anders Henriksen who completed there were many. But rocksalt com- processing and suspension formulation his Ph.D under Kingery. pounds – there just weren’t all that going on in this field, which, of course, Kingery sent me to the University of many materials of commerce. also is near and dear to the hearts of Texas at Austin to do an experiment The fact that lithium cobalt oxide ceramists. The manufacturing process with an Auger spectrometer that pro- is an ordered rocksalt compound that involves making laminates and winding fessor Harris Marcus had in his lab. I plays a central role in technology them into laminated cells. went to Austin, Texas, for three days of caught my research and teaching inter- We started thinking about how to experiments, came back, and published est. But, I went on to find out that control colloid chemistry and to do my first paper in 1981 – a paper on spinel compounds, lithium manganese something very radical, which is to try grain-boundary segregation in magne- spinel, a structure near and dear to the to get batteries to self organize. The sium oxide. That’s where I really got heart of ceramists, was also a founda- idea of this came out of research I had started in the ceramics field. tional material in lithium-ion technol- done over a number of years on surface Later, I went on to graduate school ogy. Subsequently, olivines, as well. forces with colleagues such as Roger at MIT and became an assistant profes- So it began with this interest in French at DuPont, Cannon and others, sor in 1984. It was a magical time to oxide compounds for positive-electrode who had spent many years thinking be in ceramics because I and others materials for lithium-ion batteries, and about interfacial forces. French was the were surround by luminaries, such as the recognition that these compounds first to expose me to the symbol A123 for Kingery, Rowland Cannon, Robert were certainly ceramic compounds. the Hamaker constant scaling the van Coble, Donald Uhlmann and Kent Many of the researchers in the field der Waals force between two dissimilar Bowen. were not from ceramics, and, so, I materials, 1 and 3, separated by mate- So, at some point you start becom- thought there was a lot of room for rial 2. ing interested in lithium materials? research and innovation, and tak- Of course, this was also central to In the early 1990s, after I had ing a ceramics perspective on how to much of the ceramics research around

24 American Ceramic Society Bulletin, Vol. 88, No. 9 phate, that family was used by the Vodaphone McLaren- of compounds. Mercedes team. We wanted to So, high-powered batteries have see if we could use come a very long way. The emphasis the properties in [before 2000] in lithium-ion batteries a way to produce had been toward relatively low-powered a self-organizing but long run-time devices. We thought system. In the there was an opportunity to address an end, we were able entirely different application area. Not to produce an cell phones, laptops and PDAs. But interesting lab large, high-power systems, starting with demonstration of power tools, then electric vehicles and that concept. It hybrid vehicles, plug-in vehicles, and has not proceeded, then – something I didn’t anticipate at at this point, to a all at the time – grid stabilization sys- full-scale device tems. These are really large lithium bat- development. But teries that are used to buffer short-term, in studying these high power fluctuations. materials, I and But the engine for all that was the my then-postdoc work on nanoscale olivine compounds Sung-Yoon Chung from which we thought we could A123’s aftermarket Hymotion packs can convert standard hybrid automobiles into plug-in vehicles. and then-graduate extract the high rates of charge–dis- Some plug-in hybrids can achieve over 100 miles per gallon of gasoline student Jason charge, and also take advantage of some Bloking found we other attributes of these compounds, grain boundaries and microstructure could also produce olivine compounds namely their inherently better safety development in silicon nitride and that, in a lithium-ion cell, allowed and longer life over other carbon com- other silicon ceramics that Fred Lange, charge–discharge rates that were con- pounds at the time. David Clarke Manfred Rühle and Rob sidered very high at the time. From a functional point of view Ritchie have also extensively studied. How high of discharge–charge rates are we now blurring the distinction The basic idea we started thinking are we talking about here? between a battery and a capacitor? about was how you might use surface This is a fast-moving field. Back in Yes. That Formula 1 battery delivers forces to produce batteries that would the 2001–2002 time frame, a rapidly levels of power, if you take it in terms “organize” in the sense that the cath- charging and discharging battery was of power density, which is units of watts ode and anode materials might repel considered to be one in which you per kilogram, of more than 10,000 watts themselves and spontaneously form an might be able to charge and discharge per kilogram. That is a value that peo- electrochemical junction. at a 5C rate, or one-fifth of an hour. A ple, up until now, considered achiev- In searching for materials for this twelve-minute rate then was considered able only with capactor technology. concept, we started looking at optical to be really high powered. The thing about these high-powered and dielectric properties of lithium stor- We found that with certain composi- batteries is that they also have several age compounds. It was easier to find tions of olivines, which we produced in times the energy density of the best electrode materials of high refractive a very fine powder form, with primary capacitors available. You get that power index, carbon and many metals. These particle sizes on the order of 50 nm or at a much higher energy density level. could be used as the negative electrode, less, that we could charge and discharge So, yes, that is blurring those distinc- but the hard part was finding materials at rates up to 20C or three minutes. tions. of low refractive index and also those In the lithium-ion battery field at the Did you have any training on the that could be used as the positive elec- time, this was a significant achieve- business side? When you started get- trode of the junction. ment. Just to show you how far we have ting into the issues of licensing the We started looking at olivines, come from then until now, researchers A123 technology from MIT, was that because, of all of the compounds known now talk about charge–discharge rates leap difficult? to be useful as electrode materials that of 200C. So subminute rates are pos- First of all, I want to reinforce your we screened, these had the lowest sible. starting point, which is that what is refractive indices. A123 even has a product that was special about A123 is that it was found- So, that led to some work we did used this past year in Formula 1 racing, ed, and still is based on, materials inno- looking at olivines – lithium iron phos- in which 80 percent of the theoreti- vation. It’s not only me. My cofounder, phate, lithium manganese phosphate, cal energy of the battery is discharged Bart Riley, got a degree from Cornell lithium nickel and lithium cobalt phos- in about six to eight seconds. That with David Kohlstedt in ceramics and

American Ceramic Society Bulletin, Vol. 88, No. 9 25 Yet-Ming Chiang Interview

I probably visited 30 venture capital investors, and we were really fortunate to get some top-notch investors out of this exercise. Opening those doors was certainly something I couldn’t have done myself. When you are making a pitch to a venture captial group, are you explain- ing the science and Bart is explaining the engineering and Ric is discussing the business deal? It is important to convince an investor that even the technical guys understand what the impact of this can be. You’re right, at first, but over time the boundaries blurred a great deal. Ric became an expert in the technology very quickly, and I learned a lot about A123 Systems Hybrid Ancillary Power Units can be rapidly deployed to energy grids worldwide to hybridize a power plant and provide frequency regulation and synchro- the business side from Ric and Dave, nous reserve. These units respond in milliseconds and free up power-plant capacity and it was mutually beneficial. So, I was reserves. able to help with business development as well as the technology, as was Bart. geology. Many people in ceramics know honestly, that’s the key thing. You need And it wasn’t as if the business devel- Kohlstedt and his excellent work. the technology and you need the peo- opment folks had a boundary around Riley first went to American ple and you need the financial backing. them. They knew and could recognize Superconductor after he first got his One thing I’ve learned is that you technical issues and bring them up. Ph.D. That was the first company I was shouldn’t expect and you don’t need to You end up with overlapping spheres involved in cofounding. Very quickly do it all yourself. What has made it suc- of expertise. What makes the whole after A123’s foundng, our CEO David cessful for me is finding good partners. team gel is that you are able to get Vieau joined us, and he had a back- The folks I just mentioned have really people together, learn from each other ground in mechanical engineering. performed as an excellent team. Every and blur those boundaries. From the very beginning, we had a lot one of us brought something different When it comes to licensing with an of technical leadership, particularly in to the table. As the academic research- institution like MIT, is it just a stan- the materials area. er, I wasn’t involved in A123 for my dard agreement and there aren’t really Then, the first new employees at business acumen, right? negotiations to enter into? the company were graduates from my What happened was that my other Yes, well that’s another area I had department at MIT. Even today, we cofounder, Ric Fulop, catalyzed the to become quite familiar with through probably have more materials scientists whole thing by coming to my office firsthand experience: How does tech- than any other battery company in the one day and announcing to me that he nology make it from laboratory out to world. was interested in developing a venture commercial development? In fact, I What we did was to supplement based on new battery technology, and am currently on a committee at MIT that materials expertise with people he wanted to know if I was working called Technology Transfer for the 21st who had a lot of experience and deep on anything that had that potential. Century, and we are looking at exactly knowledge in the field of batteries. What he really did when he arrived at this. What are the policies that allow That was how we built the initial tech- my office was to prompt me to start to MIT to best meet its technology-trans- nical team at A123. And that philoso- commercialize things that I might oth- fer mission of disseminating technology phy persists today. erwise have waited longer to do, waited and creating societal impact with it? On the business side, I never had for a higher level of development. Licensing is a key part of that. taken any business classes and I don’t Now, because of my day job was and A start-up company typically needs have a business degree. But my first is at MIT, Bart Riley was instrumental some exclusivity. We struck an exclu- experience being a part of a start-up, because he came on board to drive the sive license for the patents that had with American Superconductor, was laboratory development team. Ric was been generated in my lab at MIT. That a great learning experience. I started the business catalyst, who got exposure was part of the foundation of getting to learn how these ventures are pulled for us as a new venture to a broad cross a start-up going. The terms that typi- together. How you finance it. How you section of the venture community. cally go into a licensing agreement for a get the right people involved. And, In the summer of 2001, Ric, Bart and start-up include some ownership in the

26 American Ceramic Society Bulletin, Vol. 88, No. 9 company for MIT. It includes a royalty I think it was an important piece For sure, you need technology and structure and regular payments, as well rather than a turning point, because, you need customers, but building facto- as an up-front fee. So, the three pieces prior to that election, A123 was already ries is very expensive. Already, because are equity, royalty and fees. engaged with multiple automakers and of the activities with the U.S. auto- Over time, I have learned a lot about over a dozen specific automobile proj- makers, there were thoughts and plans how to structure these agreements so ects. to have manufacturing in the U.S., that it works for everybody. We were producing the Hymotion in Michigan, in particular. We were So, there isn’t one blueprint for the plug-in conversion pack. We were already manufacturing in Massachusetts licensing. There are some negotiations already producing the BAE hybrid bus at that time. So, it wasn’t as if we about how to exactly structure it in a pack for the Daimler–Chrysler Orion waited until the stimulus funds arrived way that addresses all three compo- VII bus –there currently are over 1,000 to consider manufacturing in Michigan. nents? of these on the streets in U.S. cities. Those plans started to gel well before That’s right. It really depends on the So, when the election took place we that. technology and the business. Certainly, were already in that business. We were There were very good economic and at MIT, our licensing office wants to already underway with our grid-stabili- business reasons for that. It was not a understand what the model is. There is zation product that is a 53-foot tractor- decision based on political calculus. also an attempt to create some flexibil- trailer-size type of battery, used in a But it was very helpful that there were ity, because sometimes in the beginning lineup of six or a dozen or more, and going to be loans and grants available you are not exactly sure what the model that project was with AES in southern from the stimulus package to help us will be for turning the technology into California. raise further capital and compete a business. So all of those pieces were well globally. Was having the Obama administra- underway. The stimulus package did The endorsement of the DOE tion come into office create a turning have a huge impact on the company’s through awarding of a manufacturing point for the company, given its out- ability to build manufacturing capacity grant – in which we received $249.1 look on energy and alternative trans- going forward, which is necessary to be million, the second largest of the bat- portation? a global competitor. tery manufacturing grants – was a real

American Ceramic Society Bulletin, Vol. 88, No. 9 27 Yet-Ming Chiang Interview endorsement that we were on the right ing the manufacturing process side? another MIT professor, who has been track. One thing about A123 that is quite postulating that in regard to photovol- But what led up to that was years of unique amongst other companies is that taic panel production, that assuming work with the DOE. We had multiple we really are vertically integrated. We you are using the best amalgam of projects in the automotive battery area make our own value-added electrode technologies available, the future is all with the DOE and had demonstrated materials that goes into our batteries, about making process improvements, that we could deliver. So, it wasn’t all the way up through a complete sys- and that will ultimately drive cost a situation where out of the blue we tem. down. What do you think? applied for a large grant and received it. Initally, we knew we could make the In comparison to that, let me add a There was a lot of ground laying prior value-added materials, but we started to touch of detail to what I said earlier. to that, getting smaller grants – and make cells rather than to be a materials I couldn’t agree more that process A123 delivering on those grants. producer and a materials supplier to cell improvements are very important for Didn’t some of your competitors makers. After we got in the cell-making getting costs down. make promises that they would move business, we found many customers But with battery technology, if you manufacturing in if they got the needed some expertise in designing can store more energy with less mate- award? packs – and then systems. We became a rial, that lends itself directly to reduced Well, the criteria were quite clear. pack and systems company as well. costs. The metrics are dollars-per- The phrase “shovel-ready” was used, An example of this is the grid stabi- watt for high-power applications, and and I think we were able to demon- lization product that goes by the initials dollars-per-watt-hour if it’s an energy- strate that to the DOE. HAPU, which stands for hybrid ancil- focused application, and these costs are I assume your team is quite happy lary power unit. These units produce amortized over the life of a battery. So, with the technical and business suc- two megawatts of power and have half a being able to store more energy and cess of the company to date. megawatt-hour of stored energy. It uses deliver more power with less material Yes. I look at it a little bit like it’s our workhorse cells, but also contains a over a longer lifetime, that helps to an overnight success that took eight great deal of thermal, mechanical and produce a reduction in cost, in addition long, hard years of work for every- electrical engineering. to a reduction in manufacturing costs. body involved. It’s been an incredible Regarding the future, I think that What is the status of the Michigan amount of work to get to this point. you will see progress on all fronts. plant, currently? Everyone in the company who con- The company’s approach is one of There are two primary products that tributed to that feels a great sense of being vertically integrated, and I would will come out of the planned Michigan accomplishment. say that we are working on all parts of plants. One is a cylindrical cell and one Looking ahead, how much of that vertical integration. is a prismatic cell. A123’s future depends on new tech- The reason I asked was because of nologies and how much is it in perfect- some comments from Emanuel Sachs, (Editor’s note: A fan of his com- pany’s products, Chiang gave the interview while driving his Toyota Prius equipped with one of A123’s Hymotion, an after-market plug- in module. This module allows the vehicle to get more than 100 miles per gallon while producing one-half of the

CO2 emissions of a comparable car during Chiang’s 40-mile daily com- mute to MIT.)

In a project with BAE, A123 powers over 1,000 hybrid municipal busses.

28 American Ceramic Society Bulletin, Vol. 88, No. 9