Abuse behavior of high-power, lithium-ion cells
Introduction
With increasing interest in lithium-ion batteries for automotive applications, there is a need to better understand the abuse tolerance of these batteries. For example, that fluorinated binders, such as polyvinylidene fluoride (PVDF), react exothermically with lithium is of some interest because use of fluorinated binders is desirable. Fluorinated binders perform well and are easy to use. So, the question naturally arises, do fluorinated binders contribute to thermal runaway? Or, because fluorinated binders react in the negative electrode only above 200 °C, is the reaction a consequence of thermal runaway? If so, to what extent does it contribute to the overall heat release and thus aggravate thermal runway?
The question as to the role of fluorinated binders in thermal runaway can be addressed through mathematical modeling. Through modeling, the various reactions contributing to thermal runaway can be decoupled and so clarify what reactions cause thermal runaway. To achieve this end, the model should include the chemical reactions and the mechanisms for heat conduction and dissipation. Modeling is practically indispensable in resolving this issue, since it enables a broad range of compositions and conditions to be explored, without the need for assembling and testing an extensive number of large batteries.
Here, simulation results, based on a mathematical model for abuse tolerance (oven, short-circuit, overcharge, nail and crush), are presented. The simulations are based on the thermal behavior of lithium-ion battery materials. Before presenting the model, reviews of the abuse behavior of lithium-ion batteries and modeling work are presented.
Section snippets
Survey of abuse behavior of lithium-ion batteries
The following abuse tests are widely used to characterize the abuse tolerance of lithium-ion cells.
- (a)
Oven test: This test simply involves exposing the battery, at some initial temperature, to a higher temperature. For consumer batteries, an oven temperature of 150 °C is used (UL2054).
- (b)
Short-circuit: A low resistance (<5 mΩ) is connected across the terminals of the battery. The battery may be preheated. In this test, current flows through the battery generating heat. The battery is heated internally
Survey of thermal behavior of components in lithium-ion batteries
Abuse testing of batteries gives some insight into failure mechanisms, but direct characterization of the materials that constitute the battery can lead to an understanding of what makes a battery abuse tolerant. Below, the important exothermic reactions that take place during abuse testing of lithium-ion batteries are summarized, and then the literature supporting each reaction reviewed in order to extract quantitative information useful for modeling.
The following exothermic reactions have
Estimation of reaction parameters
All the reactions are assumed to follow an Arrhenius expression for the kinetic rate constant. There is very limited information on activation energies (Ea) and frequency factors (k0). In order to estimate these values, the following procedure was followed. First, assuming all the reactions are pseudo-first-order,
Base case development
Two lithium-ion chemistries are in competition for use in hybrid electric vehicles and 42-V batteries: LiNizCo1−zO2 [22] and LiMn2O4 [21]. LiNizCo1−zO2-cathodes provide better life, but are more expensive than LiMn2O4-based cathodes. To model the abuse behavior of these chemistries, some educated guesses must be made as to the internal design of the corresponding batteries. Specifically, the dimensions of the electrodes, collectors, and separator must be known, as well as the compositions. With
Survey of thermal modeling of lithium-ion cells
A significant amount of work has been carried out to develop mathematical models for the thermal design of lithium-ion cells; only a few representative papers will be described here. Chen and Evans [35] considered the problem of heat generation in a prismatic stack of lithium-ion cells using a simple energy balance:where T is the temperature; t time; ρ the cell density; Cp the heat capacity; k a thermal conductivity; x, y, and z represent spatial dimensions;
Model development
For simplicity, a cell is modeled as a slab consisting of layers of various materials (see Fig. 3). Only one dimension is considered (x-direction), the slab is considered uniform in the y and z directions. By symmetry, only half of the cell is considered. The origin of the coordinate system is at the center of the cell. Using a control volume approach [40], mass and energy balances are carried out for each layer of the cell. The positive, separator, and negative layers are composites. The
Oven test
For a 175 °C oven test, Fig. 6 shows how the temperature profile develops in the battery over time. The figure shows the temperature distribution in one half of the battery (the other half is the same by symmetry); the core temperature is shown on the left and the can temperature on the right. As the battery heats up, the temperature is highest at the can surface and steadily decreases toward the core (see 4999.69 s). As the exothermic reactions are activated and start to release heat, the
Conclusions and recommendations
The simulations indicate that the negative electrode binder plays a relatively unimportant role in thermal runaway. The amount of binder is limited, and the binder reaction must compete with the more facile solvent reaction for lithiated carbon. Reactions involving lithiated carbon are somewhat hindered because the lithium must diffuse through the solid carbon to react at the surface. When the battery is overcharged, substantial amounts of lithium metal can form. The lithium may be in intimate
Acknowledgements
R.S. expresses appreciation for NSF Award DMI-0109141 for partial support of this work.
References (44)
- et al.
Electrochem. Commun.
(2000) - et al.
J. Power Sources
(1998) - et al.
J. Power Sources
(1999) - et al.
J. Power Sources
(2001) - et al.
Solid State Ionics
(1994) - et al.
J. Power Sources
(1999) - et al.
J. Power Sources
(1999) - et al.
J. Power Sources
(2001) - et al.
J. Power Sources
(1999) J. Power Sources
(2001)