Summary

With the invention of new electrode materials and reduced manufacturing costs, electrochemical energy storage has become firmly established as a power source for electric vehicles (EVs). The safety of the battery pack requires thermal management, health monitoring, and external shock protection design for the battery pack. To better understand how batteries respond to external mechanical loads, predictive simulation tools can be improved to help design more advanced and safe battery packs and lightweight metal enclosures.

In 2017, the "Journal of Power Sources" published a research work by the Oak Ridge National Laboratory's Computational Science and Engineering Division on the strain distribution and failure modes of lithium-ion battery polymer separators under biaxial loading.

This article studies the biaxial tensile deformation of polymer separators of lithium-ion batteries, uses DIC technology to measure the strain of the separators on-site, and uses the finite element model to discuss the consistency between the test results and the finite element results.

Test device: 100 lbs load cell, circular sample is placed between stainless steel 304L flanges to avoid being torn, the flanges are fixed with nuts, polished steel balls realize biaxial tension, three ball diameters are 1 inch, 2 inches, 2.5 inches, polished and sprayed with anti-friction coating, the steel ball is supported on the concave surface of the coupler, the system has no rigid connection to avoid bending and torsion, the ball speed is 0.008 inches/second.

Microstructural level

Fig.2.Microstructure of separator Celgard 2325: a) before deformation, b) after deformation; Celgard 2075: c) before deformation, d) failure mode after deformation

The separator is composed of longitudinal (MD) fibers and transverse (TD) polymer thick sheets. The deformed microstructure is shown in Fig. 2b/2d: lamellar extension and small cracks occur simultaneously, new fibers are formed by fiber elongation and subsequent sheet separation, Lamellar splitting and fiber pullout results.

Experiments found that due to the accumulation of longitudinal strain, both materials will undergo longitudinal stretching, especially Celgard 2325 is more obvious. Celgard 2325 may show cracks in the necked region when deformed to higher strains and eventually fail, while Celgard 2075 shows no stretch marks and fails without necking. Figure 4 shows that the translucent zone of Celgard 2325 developed significantly, probably due to the supporting role of the intermediate PE layer. Both diaphragms studied failed with straight cracks along the longitudinal direction, Celgard 2325 had significant longitudinal stretch and "necked" resulting in unstable deformation. The difference in deformation and failure behavior of the two diaphragms is that Celgard 2075 has no stretch marks, and Celgard 2325 may have higher deformation and cracks in the necked area. Figure 5 compares the strain fields of the two diaphragms before the first appearance of the translucent diaphragm in Celgard 2325. The results show that the critical principal strains of the two diaphragms are very close regardless of the size of the spheres. For the Celgard 2325 diaphragm, the necking strain is 0.34±0.05, and the failure strain is 0.79±0.34; while for the Celgard 2075 diaphragm, the principal strain at failure is 0.43±0.069.

Failure mode of longitudinal diaphragm under biaxial load deformation mode (a) Celgard 2325, 50.8mm sphere; (b) Celgard 2075, 25.4mm sphere

Strain concentration and formation of transparent tensile bands in Celgard 2325 just before failure

Principal strain distributions of (a,b) Celgard 2075 under biaxial loading: and (c,d) Celgard 2325

Finite element model

 In this paper, the membranous shell elements in LS Dyna are used to build the finite element mesh of the diaphragm with different thickness (0.025mm for Celgard 2325 and 0.020mm for Celgard 2075). The diaphragm radius is 38mm and the nodes are constrained in all directions around them. Three spheres with different radii were used in the simulation to represent the steel balls in the test. The load-displacement curve is shown in Figure 7. The results show that the deformation of the diaphragm is related to the diameter of the sphere. Since Celgard 2325 has a three-layer diaphragm with a higher thickness, the force generated by the deformation of Celgard 2325 is about twice that of stretching Celgard 2075. Figure 8 shows the strain distribution diagrams of two diaphragms under the action of 12.7mm and 25.4mm radius indenters. When the spherical indenter was vertically displaced by 15 mm, the transparent stretch region (neck) in the diaphragm appeared for the first time, and the predicted strain value was close to the measured strain value at this time. Due to the anisotropic properties of the diaphragm, the main strains in the figure show an elliptical profile, and the positions of the strain maxima arranged along the longitudinal direction are in good agreement with the experimental observations.

Diaphragm meshes used in biaxial deformation with spheres of different radii

Experimental and calculated force-displacement curves under biaxial test a) Celgard 2325; b) Celgard 2075

Biaxial deformation (a, b) Calculated distribution of the first principal strain in Celgard 2075; (c, d) Celgard 2325

Summary

The article carried out biaxial tensile experiments on two kinds of diaphragms, Celgard 2325 and Celgard 2075, and found that the measured value of the first principal strain at the critical point of diaphragm failure was close to 0.34 (Celgard 2325) and 0.43 (Celgard 2075), regardless of the diameter of the test sphere. The results of finite element simulations using the anisotropic mechanical properties of the diaphragm as input to the model show good agreement with the experimentally measured critical strain and the location of the strain maximum at which the diaphragm eventually fails.