Effectiveness of the United States Advanced Battery Consortium as a Government-Industry Partnership (1998)

This chapter provides a review of the technical progress made in USABC battery technology R&D programs through the end of 1997. The committee's review addresses work done in Phase I, commitments for Phase II, and work in progress through 1997. The chapter is organized by battery technologies, namely, Ni/MH (nickel metal hydride), lithium, and fused-salt battery systems. The committee assessed the potential of each technology for EVs on the basis of progress made by the USABC and the technical goals and objectives discussed in Chapter 3. The committee also identified the limitations of each technology.

The Ni/MH battery uses a nickel positive electrode (developed a hundred years ago) and a hydrogen-absorbing alloy negative electrode (developed in the 1970s). The packaging and most of the manufacturing processes are similar to the nickel-cadmium (Ni/Cd) battery. Metal hydrides used for the negative electrode include AB₅-type alloys based on rare earths and nickel, which are being developed by SAFT, and AB ₆₅₂-type alloys based on nickel and vanadium, which are being developed by GM-Ovonic. The metal hydride is an environmentally friendly replacement for the toxic cadmium in the Ni/Cd battery; the Ni/MH battery also has a higher specific energy than the Ni/Cd battery. Small Ni/MH batteries are used extensively in portable consumer electronics, such as cellular phones, laptop computers, and VCRs. Technical information on the three USABC Ni/MH battery programs is summarized in Table 4-1.

In May 1992, the USABC awarded Ovonic Battery Company a contract to develop a Ni/MH battery for EVs based on Ovonic's proprietary metal hydride AB₂-type alloy. The goal was to meet the CARB mandate for the commercial introduction of EVs (with subsidies) in 1998, i.e., the system was to be developed as a midterm battery. In 1994, Ovonic Battery Company and GM entered into a joint venture, GMO (GM-Ovonic), to manufacture the battery developed under the USABC contract. The original contract called for a three-year program consisting of three phases¹ with increasing cell performance and module and battery-pack deliverables. The deliverables were to undergo testing at GMO, at Argonne National Laboratory (ANL) under a CRADA, and in EVs developed separately by Chrysler, Ford, and GM.

Prior to the USABC contract, Ovonic had developed 25-Ah to 35-Ah prismatic cells that could deliver more than 600 cycles and had a specific energy of 54 Wh/kg. The purpose of the first phase of the USABC project was to duplicate the performance of these cells in multicell strings. The second and third phases focused on developing larger cells with higher specific energy. Deliverables during this program were multicell modules and battery packs, with 4 to 10 modules, for both laboratory-scale testing and vehicle testing. The modules and battery packs were expected to meet the midterm goals of the USABC. GMO has produced individual cells that meet the midterm goals, but their specific energy is lower at the module and battery-pack levels.

The modules and battery packs are being tested at ANL and in EVs by the three major U.S. car companies. The close interaction between the car companies and GMO in the early stages of battery development has led to improvements in the design (e.g., pressure release vents and terminal seals) and packaging of the cells and modules and to a better understanding of systems-level integration. The best packs deliver 450 to 600 cycles, which is close to the midterm requirements. The battery packs lose their power density with cycling and, therefore, do not meet the USABC calendar life requirements. The early EV tests raised questions about thermal control of the battery, so GMO made some preliminary tests of cell performance at different temperatures. The results of these tests showed that the modules will require active cooling, which will affect the specific energy of the battery. Preliminary failure modes and effects analysis (FMEA), safety, and charge control issues have also been addressed.

In Phase I of the USABC program, GMO has increased the specific energy of Ni/MH cells from 50 Wh/kg to 70 Wh/kg. The increase in specific energy and

performance of the battery resulted from significant changes in the electrode materials, including changes in the chemical composition and manufacturing processes for the positive and negative electrodes. These changes were the result of materials research conducted parallel to the development of larger cells. Materials changes in the cell designs initially caused premature failures and a loss in power density after cycling. After several iterations, however, GMO was able to rectify most of these problems.

Cost analyses were conducted throughout the program. The final cost estimate for GMO's Ni/MH battery for production of 20,000 packs per year is more than 1.5 times the midterm goal of $150/kWh. These values are close to the estimates by SAFT for its Ni/MH battery and, in both cases, a large part of the cost (60 percent) is for materials. High cost is obviously a major problem for the acceptance of Ni/MH technology in EVs. The cost of a product is very difficult to estimate when significant materials and design changes are still being made, and the committee has reservations about the validity of the cost analysis for Ni/MH presented at the first committee meeting (Rauhe, 1997). The $150/kWh target is the selling price for the battery, whereas the high cost estimate is for the production cost of the battery. For example, the selling price of small Ni/MH batteries for portable applications is currently higher than $600/kWh.

In June 1994, a 16-month Phase IV was added to the Ovonic Battery Company contract. The program was amended to include significant work on electrode materials to reduce costs and to increase the specific energy of the battery beyond the midterm goals. This was high-risk research, which was very different from the evolutionary improvements made in Phases I, II, and III of the Ovonic program. However, Phase IV did not result in significant reductions in cost.

In December 1992, the USABC entered into a contract with SAFT America, Inc., of Cockeysville, Maryland, to develop Ni/MH batteries based on an AB₅ alloy for a midterm EV battery.² The original contract lasted through March 1996. In October 1994, additional funding accelerated parts of the program (i.e., delivery of the first battery pack by the end of 1994) and increased the number of deliverable 40 kWh battery packs to five.

Since the 1970s, SAFT has built Ni/Cd cylindrical and prismatic cells for portable applications, such as laptop computers, camcorders, and power tools. SAFT used a hydrogen-absorbing AB₅-type alloy in the negative electrode of NiMH batteries throughout Phase I of the USABC program. At SAFT, the Phase I USABC program addressed the design of cells and battery packs, cost, FMEA, safety analysis, the development of diagnostic criteria, the development of cells

with higher specific energy, and the development of high energy density modules. Deliverables included cells and modules with six 40-kWh battery packs to be delivered at the end of the program.

SAFT has made improvements in both the positive Ni electrode and the negative hydride electrode and has produced cells with a specific energy of 70 Wh/kg. Five cells are connected in series (6 V modules) to produce a maintenance-free monoblock with active water cooling and a specific energy of 65 Wh/kg. As many as 56 modules are required to form a battery pack. The modules meet all of the midterm performance criteria except specific energy. Data on improved AB₅ alloys suggest that energy density per module can be increased to approximately 70 Wh/kg. Independent tests have been initiated by Ford, Chrysler, GM, DOE, the Idaho National Engineering and Environmental Laboratory, and EPRI on the battery packs delivered to the USABC, but the cycle life and calendar life of the SAFT battery have not been determined.

The cost of the SAFT and GMO Ni/MH batteries is estimated to be at least 65 percent higher than the midterm goal. The estimate is for 25,000 battery packs per year. SAFT believes that further improvements will bring the cost closer to the midterm goal of $150/kWh. SAFT has presented a detailed plan for the pilot-plant production of battery packs.

In 1994, the USABC awarded Yardney Technical Products a contract to develop a low-cost, fiber-based, pasted Ni electrode for EVs. The goals of the program were to develop a high-capacity electrode at a cost of $43.60/kWh with no more than a 20 percent loss of capacity over 1,000 dynamic stress test (DST) cycles.³ At the end of the project, Yardney claimed to have produced the high-capacity electrodes with an estimated cost of $56/kWh at production rates of 20,000 battery packs per year and a 7.5 percent loss of capacity after 600 cycles. The electrodes were provided to ANL and the two Ni/MH battery developers, GMO and SAFT, for testing. Both battery companies found that the Yardney electrodes delivered significantly lower capacity than the positive electrodes that had been developed in house. As a result, the USABC terminated the Yardney project.

Lithium metal, because of its low equivalent weight⁴ and its electrode potential, is an attractive candidate material for the negative electrode in high energy

density batteries. The positive electrode consists of intercalation compounds,⁵ such as sulfides and oxides of transition elements, that can reversibly intercalate lithium ions. During the initial development of lithium rechargeable batteries in the 1960s and 1970s, organic solvents containing lithium salts were used as electrolytes. However, the cells had serious safety problems because of the high chemical reactivity of lithium metal in liquid electrolytes. Two solutions to this problem are being investigated by the USABC. The 3M/Hydro-Québec lithium-polymer battery replaces the liquid electrolyte with a solid polymer that contains a lithium salt, which significantly reduces the tendency for reaction between the lithium and the electrolyte. The second approach, in which the lithium metal is replaced by an intercalation compound that has a sufficiently negative electrode potential, such as graphite, is used in the Duracell/VARTA lithium-ion battery. In this battery, no lithium metal is present, and the lithium ion shuttles between the two intercalation compounds. Lithium-ion batteries, also called ''rocking chair'' batteries, are used extensively in high-end portable consumer products. The W.R. Grace lithium-ion-polymer battery attempted to combine the benefits of the lithium-ion and polymer approaches. Technical information on the USABC's three lithium battery programs is summarized in Table 4-2.

A team⁶ led by W.R. Grace was funded beginning in April 1993 for the development of an organic electrolyte-polymer lithium-ion system with a carbon anode and a manganese oxide cathode. Both bipolar and monopolar configurations were developed; however, the bipolar design was abandoned in favor of the monopolar design midway through the project. The system required the development of a new battery type, including the development of new materials, before it could reach the prototype phase. The goal of the project was to provide a proof-of-concept system, including the production of a 20 kWh battery. Based on the

state of the technology, this goal was modified to provide only a demonstration of the viability of the technology in a full-scale building-block module.

Models were developed to facilitate cell and battery design and to predict thermal behavior. Safety and reliability issues were addressed through thermal modeling and the development of a preliminary system FMEA. A study of production costs was also partially completed. By the end of the program in 1996, the three-year program had yielded laboratory prototype cells and a pilot-plant facility, which had produced more than 4,500 meters of electrode material. Scale-up of the material width to a size required to build cells and modules for EV application was progressing.

Although significant progress was made, the goals were not met. Cost, cycle life, energy and power density, and electrical controls were still outstanding issues. Significant development of the materials technology would have been necessary before the long-term goals could be met. For business reasons, W.R. Grace decided not to pursue further development of this technology. The USABC has indicated that the decision not to pursue Phase II was a joint decision by the contractor and the USABC.

The Duracell/VARTA lithium-ion battery proposal was solicited and funded in January 1995, near the end of Phase I. The objective was to develop a midterm battery, although, in the opinion of the USABC, the technology could eventually approach the long-term goals. Unlike the 3M projects, this lithium-ion technology has a liquid organic electrolyte. (The USABC did not initially pursue a liquid organic electrolyte system, but in 1995 liquid electrolyte systems became attractive for use in hybrid vehicles.) The Duracell/VARTA battery is of interest to the Partnership for a New Generation of Vehicles (PNGV) program as well as the ZEV targeted in the USABC program.

The comprehensive Duracell/VARTA proposal reflected the long-term experience of both Duracell and VARTA in designing and marketing battery products for a variety of challenging applications. Planned activities included the development of electrical monitoring and control, thermal management, and safety testing and simulation, as well as FMEA. One of the deliverables was a proposal for a manufacturing process that incorporated statistical process controls. Cost analysis, disposal, and recycling were also addressed in the plan. Large prismatic cells with cobalt oxide cathodes had been developed independently by Duracell/VARTA prior to their USABC proposal.

Under the USABC contract, a lithium-ion battery with a manganese oxide cathode was scheduled to be developed. The initial program was redirected to incorporate further abuse testing, including a possible change in cell design and modified materials. The first phase of the program was completed in January 1997.

At the end of 1996, a Phase II contract was established that extended the program for 24 months. Phase II focuses on the incorporation of nine previously identified concepts to make EV-sized modules safe and includes hardware assembly, testing, and optimization. Lowering costs by optimizing materials processing and cell assembly was also planned. Testing to date suggests that the midterm performance criteria will be met, although cost and life goals have yet to be demonstrated. The Duracell/VARTA program has conducted effective module and systems engineering tests for safety and reliability.

The 3M/Hydro-Québec project is the only one presently being funded to meet the USABC long-term performance criteria. At the time the contract was signed in December 1993, the basic technology was already reasonably mature. Hydro-Québec had been working on its development for approximately 15 years, at a cost of $40 million (Canadian), for energy storage applications. The battery has a lithium anode, a solid polymer electrolyte, and a vanadium oxide cathode and operates over a temperature range of 60 to 100°C.

The USABC considered the lithium polymer technology to be the best candidate for meeting the long-term goals by 2003. The USABC provided a significant boost to its development, not only by providing additional funding and a possible market, but also by helping to fill in several gaps in the development program. Teaming up 3M (which had submitted a separate proposal), ANL, and Hydro-Québec in a three-way CRADA provided a number of benefits. The participation of 3M brought world-class R&D and manufacturing know-how in thin-film polymers to the partnership, as well as addressing the USABC goal of developing a U.S. battery manufacturing base. ANL was brought into the project to provide expertise in battery design, modeling, and testing, as well as materials development for advanced positive electrodes.⁷

The USABC also brought battery technologists together with EV experts to provide expertise in systems integration and testing. Although this expertise was important to all of the USABC projects, it was crucial to 3M and Hydro-Québec, which, unlike Johnson Controls and VARTA, had no previous experience designing batteries for the automotive industry. The committee noted that the 3M project leader was very positive about working with the automotive companies through the USABC.

The goals of the two-year Phase I segment of the 3M/Hydro-Québec project were to meet or exceed the USABC's midterm performance criteria; to demonstrate design feasibility for lithium polymer battery technology; to explore the properties and limitations of advanced materials used in the particular system

under development; and to evaluate configurations that could meet USABC's long-term goals. In the committee's view, the project was well planned, which was apparent in the technology transfer from Hydro-Québec to 3M early in the project. Manufacturing facilities were built, and the development of manufacturing processes was begun immediately at both 3M and Argo-Tech, a subsidiary of Hydro-Québec. Tasks were defined to develop models for battery performance and cost and to address safety and reliability, electrical control, thermal management, and systems integration issues. The management practices and statistical design of experiments were also good. The project leader at 3M has extensive experience in managing large government programs, which, the committee believes, contributed substantially to the success of the project.

By the end of Phase I, a 100-Wh lithium polymer battery composed of a parallel arrangement of five 20-Wh prismatic cells was delivered. Projected costs for the battery were significantly higher than the interim commercialization target. In addition, the battery cell balance and control strategy had been developed, and the preliminary battery design had been completed. Significant safety testing had also been completed at the 100-Wh cell level. The performance of the battery was projected to exceed the midterm goals. The most significant issues that still needed to be addressed to meet the interim commercialization goals were cycle life, rate of discharge, and cost. However, the battery could not meet power requirements at temperatures below 40°C because the solid polymer electrolyte has low ionic conductivity. A reliable thermal management system must be developed for vehicle start-up after the battery has cooled down.

The objective of Phase IIA is to demonstrate the viability of the lithium polymer battery technology in meeting the USABC interim commercialization performance goals at the 1.7-kWh modular battery level. In June 1997, the USABC awarded additional funding for reducing cost and improving performance. This Phase IIB, which will last for two years, will culminate in the first production series of full-sized battery packs.

At the beginning of Phase IIA, electrode and electrolyte materials were still being developed and selected to establish stable performance. Two 12-V, 760-Wh submodules were being delivered. Scale-up, delivery, and safety testing of the technology with 1.7-kWh modules and minipacks were scheduled for June 1997; another six minipacks were scheduled to be delivered in December 1997. The process for making the cathode/solid polymer electrolyte laminate is now being scaled up. At the end of 1998, if Phase IIB has been successful and the original equipment manufacturers of EVs make a commitment to purchase a given number of batteries, 3M and Hydro-Québec plan to expand their manufacturing capability at their own expense to support a market launch of the lithium-polymer technology. A final design freeze is scheduled for July 1999. A total of nearly $85 million will have been invested by the USABC, 3M, and Hydro-Québec to bring the lithium polymer battery concept to the prototype level (Letourneau et al., 1997).

Based on the most recent quarterly progress report (April through June 1997), the committee determined that the 3M/Hydro-Québec project appears to be on target to meet its deliverables. The program has made significant progress in the development of manufacturing capabilities and has assembled and tested an impressive number of prototypes. The program appears to be well managed, and the contractors appear to be committed to the program.

Two fused-salt battery systems were chosen for development by the USABC: sodium-sulfur (Na/S) and lithium-iron disulfide (Li/FeS₂). Both are elevated temperature systems. The Na/S system operates at a nominal temperature of 325°C, and the Li/FeS₂ system operates in the 450 to 500°C range, depending on the composition of the fusedsalt. The Na/S system uses two ionic conductors in series: a ceramic solid electrolyte system and a sodium polysulfide salt, which is a liquid at operating temperatures. The ceramic is a sodium-ion conductor that mechanically separates the positive and negative reactants and also limits the conductivity to sodium ions only, thus preventing cell self-discharge.

In August 1993, a contract was signed with the American office of the Silent Power group, which was owned at that time by the German RWE group. RWE had purchased the background experience of the British company, Chloride Silent Power, Ltd., which had been working on the Na/S system for more than 20 years. The USABC contract was structured to exclude all cell development and to focus on generic battery designs, features, and processes that could be applied to any candidate Na/S technology. The Na/S technology was also being investigated by other groups in Germany, Japan, and China.

The Silent Power project addressed the areas of thermal management, mechanical assembly (especially reliability and cost), systems integration, and controls. Various approaches to cooling were studied, including cooling by air or liquid (an oil) and by varying hydrogen pressure, referred to as variable conductance. Variable conductance was investigated at the National Renewable Energy Laboratory. Liquid cooling was studied, with the variable conductance as a backup.

A major goal of the American office of Silent Power was to provide an improved design of the Na/S battery that could be used for EVs, and the planar module design did meet the USABC requirements and passed several tests. An 840-cell battery was subjected to extreme failure testing, including dropping the battery onto a steel post, which severely distorted the battery enclosure and caused the noncatastrophic failure of 21 cells. The battery was also subjected to prolonged vibration. Cost analyses showed that the midterm USABC price goals could be met, given adequate levels of production.

The automotive Na/S battery designed for the USABC involved a new cell design that was being developed by the parent company, Silent Power RWE. At the end of the USABC contract, the cells had not yet demonstrated required life expectations. For this reason, the USABC decided to terminate the project. Under the USABC contract, the cell had been redesigned and tested by Silent Power in England and Germany; those results were made available to the USABC, although Silent Power RWE retained the rights to the technology.

This USABC project was difficult to assess because most of the requisite data were supplied by Silent Power RWE. The committee was not provided with the program design or the results of prolonged testing of the new cell design. The American office of Silent Power appears to have directed its efforts toward crucial design areas and to have made significant progress toward meeting its objectives.

A contract between the USABC and SAFT America, Inc., for the development of an advanced, high-temperature Li/FeS₂⁸ battery for EV propulsion was signed in December 1992. A CRADA was established between SAFT and ANL, which has a long history of working with the Li/FeS₂ system. SAFT took the lead in module development, and ANL assumed responsibility for technical support.

Technical issues for the Li/FeS₂ system include the loss of capacity caused by the migration of reactants through the separator, the composition of the separator, the optimum composition of the fused-salt, and the construction of long-lived corrosion-resistant seals. Solutions to these technical problems are likely to increase the cost of the battery.

Studies of both the positive iron disulfide (FeS₂) electrode and the negative lithium alloy electrode were conducted, and the performance of both was improved. The paste electrolytic separator, based on magnesium oxide, was also studied and improved. The fused-salt electrolyte itself was investigated for various compositions, and an optimum ternary composition involving lithium ion as the only cation was found. This composition lowered the melting and operating temperature and increased ionic conductivity.

A good deal of effort was directed toward improving seals and cases. At the elevated operating temperatures of the battery, the salt components of the cell tend to be aggressive corrosion agents, which drastically shortened the lifetimes of the seals and cells. Studies were done on the packaging design of cells and

modules and the temperature management of the resulting package. Studies were also done on managing the consequences of overcharging and overdischarging.

At the conclusion of the contract, the cells were capable of 200 Wh/kg and nearly 500 W/kg. The project was mutually terminated because of shortfalls in cycle life, calendar life, and cost.

Finding. The USABC brought together battery technologists and experts in EVs, thereby providing a valuable systems integration and testing focus to the battery development programs.

Finding 1. Significant improvements have been made in Ni/MH battery technology in the programs coordinated by the USABC, particularly in developing large cells and batteries for EVs. Furthermore, USABC projects have led to a better understanding of design, manufacturing processes, and cost issues for Ni/MH systems and have established processes for the pilot production of electrodes, cells, and modules. However, significant work on battery systems-level integration and evaluation remains to be done before the technology will be roadworthy.

Finding 2. Ni/MH technology meets the midterm targets for power density and volumetric energy density but does not meet the most critical midterm criteria for specific energy and cost. The calendar life of Ni/MH batteries has yet to be determined.

Finding 3. Because the cost of the Ni/MH batteries is very high, the USABC has decided to focus on reducing the cost in Phase II. It remains to be determined whether the costs can be reduced enough to meet the USABC's interim commercialization goals.

Finding 4. The USABC's decision to develop two competing Ni/MH technologies based on AB₂-and AB₅-type hydrogen absorbing alloys (GMO and SAFT technologies, respectively) provided a good insurance policy in the event that either of the technologies failed to meet program objectives.

Finding 5. The Yardney program to develop a low-cost nickel electrode suffered by not interacting with battery manufacturers early in the program. Because electrode and cell designs could not be satisfactorily integrated, the performance of the battery was disappointing.

Finding 1. Substantial progress has been made on lithium batteries, although all systems failed to meet their cost objectives. The 3M/Hydro-Québec lithium polymer battery and the Duracell/VARTA battery both met the midterm performance requirements, although cycle and calendar life have not been demonstrated yet because of the long test times. Preliminary results suggest that they may not meet the long-term goals.

Finding 2. The Duracell/VARTA battery may be able to meet the midterm goals, although not in the midterm time frame, and may offer opportunities for further development to meet USABC interim commercialization goals; there is a separate program for meeting PNGV requirements for hybrid vehicles. Some safety issues remain to be addressed.

Finding 3. The 3M/Hydro-Québec battery is projected to come close to meeting the interim commercialization goals and comes closest of all the technologies to meeting the long-term goals. Further work will have to be done on the recycling or disposal of spent batteries, safety testing, and performance testing under extreme temperatures and stressful-use conditions. A reliable thermal management system will have to be developed for vehicle start-up after the battery has cooled down.

Finding 4. Final design freezes for the midterm Duracell/VARTA lithium-ion battery and the long-term 3M/Hydro-Québec lithium polymer battery are planned for mid-1998.

Finding 5. The W.R. Grace team made progress in developing its lithium-ion polymer technology from an embryonic stage, as the USABC requested. Nevertheless, the technology was not ready for full-scale development within the USABC time frame.

Finding 6. The 3M/Hydro-Québec lithium-polymer battery program has made substantial technical progress in the context of the USABC goals and time frame.

Finding. The SAFT-ANL project to develop a Li/FeS₂ battery was well organized, but not enough emphasis was placed on known problem areas for high-temperature fused-salt systems, namely, cell cycle life and calendar life.