Energy Storage for Automotive Propulsion, Part II

Advanced Battery Technology,  May 2006  by Howard, W F

• E-mail
• Print
• Link

(Part I of this article appeared in April 2006 ABT. )

Most Li-ion batteries found in present EV/HEVs incorporate cathodes with LiCoO^sub 2^ derivatives and anodes of graphitized carbon: these represent the established production technology. Even the newest generation units, with low Co content (reducing price) and stabilizing Al (less chance for thermal runaway), require expensive (and heavy) safety circuitry. And the cathode material is still the high-cost component: reports by Argonne National Lab and other groups cite the cathode intercalant as >20% of the battery cost. This percentage will drop as Co content is reduced or eliminated.

Spinel was for many years considered a disruptive technology against LiCoO^sub 2^, and oceans of R&D money broke against the technological cliffs preventing spinel's adoption into Li-ion batteries. Modified LiMn^sub 2^O^sub 4^ has much going for it: low cost, stable raw material sources, safe operation, and only minor health and environmental issues. But the bottom line is performance, and this is where spinel foundered. Capacity is relatively low and operating life generally less than 1000 cycles, non-competitive with LiCoO^sub 2^. The source of this latter problem was recognized quickly: Mn^sup +3^ in spinel is subject to disproportionation, a behavior exacerbated by trace acid in the electrolyte. Mn^sup +2^ and Mn^sup +4^ thus produced interfere with cell operation, causing unacceptable capacity fade.

Rapid deep-discharge fade and low capacity are incompatible with EV batteries, which require extended lifetimes, but the excellent rate capability of spinel suggests application in HEV cells. For example, LG Chemical has a spinel/Li polymer battery producing 5kW/kg. Although Nissan has ventured into spinel-based cells, it appears that the concerns over limited cycle life, even in pulse usage, restrict spinel's automotive battery future. Even so, Nissan projects power and energy ratings of 3.5kW/kg and SkWh/ kg, respectively, by 2008. PolyStor and Panasonic determined that physical mixtures of spinel and LiCoO^sub 2^ derivatives yield a cathode with the best properties of both materials, but the move away from Co has kept this concept from appearing in HEVs.

Lithium metal polymer (LMP) batteries from Avestor (Canada) and Bolloré (France) incorporate Li^sub 1.3^V^sub 3^O^sub 8^ cathode, PEO-based electrolyte, and Li metal anode. These warm batteries run at 60-80°C and are best suited for EVs (120-140Wh/kg, only 250W/kg). Bolloré, with manufacturing by its BatScap subsidiary, projects their ~200kg battery pack will sell for

Consider a rocking chair battery with LiCoO^sub 2^ (LCO) cathode and Li^sub 4^Ti^sub 5^O^sub 12^ (LTO) anode. Although both electrode materials are very fade-resistant, cell voltage is low (~2.3V) and capacity is moderate, probably ruling out usage in EVs. Excellent high rate performance (full charge in three minutes), as reported by Amatucci, ensures that this technology will be thoroughly evaluated for HEV use. Toshiba introduced a similar battery in 2005, claiming only 1% fade per 1000 cycles. This extremely low fade leads to speculation that the cathode incorporates LiFePO^sub 4^, a combination without overcharge-overdischarge problems, a key feature in the safety-conscious auto industry. Production of nano-LTO continues to be a question: can the process be scaled to tons/week output and the price reduced to competitive levels?

Sony developed a nano-structured SnCoC alloy anode (also described by J. Dahn) paired with Li(Co,Mn,Ni)0^sub 2^ cathodes in 14430 cells that charge to 4.6 V and produce 20% more capacity than LCO-based units. While this anode material has 100% expansion with Li uptake, it does not fragment on cycling and discharges in the 0-0.4V range, thus allowing greater energy per cell than with LTO.

LiFePO^sub 4^ in a conventional Li-ion cell with carbon anodes yields a relatively low-energy system, but one with extreme stability. Not only is the risk of thermal runaway overcome, a trait common to phosphates, but deep-discharge cycle life may exceed 2000 cycles. Development efforts with LiFePO^sub 4^ are geared toward laptop-sized batteries, although Valence Technology announced their intention to target automotive applications (possibly with Li^sub 3^V^sub 2^(PO^sub 4^)^sub 3^, which cycles at 3.5-4.6V and has greater capacity - 195mAh/g - than LiFePO^sub 4^). There are still difficulties with LiFePO^sub 4^ production, and until the process variables are controlled, cathode material cost and quality will be problematic.

Finally, a few words about Li-ion electrolytes. Typically the electrolyte is a 1M solution of LiPF^sub 6^ in organic carbonates. This is a very reactive salt that hydrolyzes readily and contains trace HF, but it is still the preferred candidate for the job. Other Li salts, such as LiBF^sub 4^, LiCIO^sub 4^, and imide derivatives, cannot withstand the high potential or meet the conductivity demands of 4V Li-ion cells. While LiAsF^sub 6^ has similar electrochemical characteristics as LiPF^sub 6^, and is very stable and acid-free, the spectre of arsenic has prevented its acceptance by the battery industry. Lithium bis(oxalato) borate (LiBoB) is also stable at high voltage, but production difficulties and low conductivity have restricted this salt's market entry. Gel electrolytes are electrochemically almost identical to salt solutions, and provide some advantage in battery construction. Adoption of these formulations has been slowed due to cost considerations.