Particle Compression and Conductivity in Li-Ion Anodes with Graphite Additives C.-W. Wang, a Y.-B. Yi, a A. M. Sastry, a,b, * ,z J. Shim, c, * and K. A. Striebel c, * a Department of Mechanical Engineering and b Department of Biomedical Engineering, University of Michigan, Ann Arbor, Michigan 48109-2125, USA c Lawrence Berkeley National Laboratory, Environmental Energy Technologies Division, Berkeley, California 94720, USA We performed coupled theoretical/experimental studies on Li-ion cells to quantify reductions in anode resistivity and/or contact resistance between the matrix and the current collector with the addition of amorphous carbon coatings and anode compression. We also aimed to identify microstructural changes in constituent particles due to anode compression, using models of permeable- impermeable coatings of graphite particles. We studied three anode materials, SL-20, GDR-6 6 wt % amorphous carbon coating, and GDR-14 14 wt % amorphous carbon coating. Four compression conditions 0, 100, 200, and 300 kg/cm 2 were examined. Experimental results indicated that electrical resistivities for unpressed materials were reduced with addition of amorphous carbon coating for unpressed materials: SL-20 GDR-6 GDR-14 ). Contact resistances were reduced for SL-20 anodes by the appli- cation of pressure. Overall, the two-dimensional 2Dimpermeable particle mathematical model provided reasonable agreement with the experiments for SL-20 and GDR-6 materials, indicating that coatings remain intact for these materials even at moderate pressures 100 and 200 kg/cm 2 . Conductivities of SL-20 and GDR-6 anodes exposed to the highest pressure 300 kg/cm 2 fell short of model predictions, suggesting particle breakage. For the GDR-14 graphite, both 2D models underestimated conductivity for all processing conditions. We conclude that the 2D simulation approach is useful in determining the state of coating. © 2004 The Electrochemical Society. DOI: 10.1149/1.1783909All rights reserved. Manuscript received August 6, 2003. Available electronically August 18, 2004. The use of lithium metal, a powerful reducing element, with a strong oxidant e.g.,V 2 O 5 , MnO 2 , LiNiO 2 , or LiCoO 2 ), allows the realization of high voltage, high energy density cells. But safety and cycle life problems due to the dendritic morphology of the charge- deposited lithium metal have restricted the use of Li as a composite matrix. Recent research has thus focused on improving energy den- sity and cycle life of lithium-ion cells by using graphite, carbon- coated graphite, tin oxide, and intermetallic compounds as Li inter- calative host materials. 1-3 Specifically, high electronic conductivity is critically important in Li intercalative host materials. 4-6 But the high electrical resistivity of graphite matrix materials relative to Li metal in the anode neces- sitates the use of additives to improve conductivity. Recently, our group has studied improvement of conductivity through the addition of conductive particles. 7 This approach was guided by classical work on effective and percolative properties of systems of high as- pect ratio particles. 8-11 Specifically, we have demonstrated that mod- erate increases in the particle aspect ratio particle length/particle diameter, L / d ) for low-density materials provide significant im- provements in electrical conductivity, 7 due to the dramatic reduction in the percolation point. 12 The use of percolation models allows the prediction of both ther- mal and electrical conductivity vs. density and particle shape. High thermal conductivities of both anodes and cathodes have been shown to be important in preventing thermal runaway, when cells are operated at high temperature. Recently, Maleki and co-workers 13 studied the thermal conductivity of anodes comprised of synthetic graphites of various particle sizes, various fractions of polyvi- nylidene difluoride PVDFbinder, and carbon black. Several levels of compression were also used. They found that the highest thermal conductivity was achieved using the largest graphite particles 75 m, the lowest carbon black content 5%, and the highest pressure 566 kg/cm 2 . They also observed a doubling of thermal conductiv- ity at room temperature 27°Cwith an increase of pressure from 250 to 575 kg/cm 2 . Their results are generally consistent with the percolation model predictions; higher conductivity can be achieved with higher volume fractions of particles, produced by compression of the electrode. However, high compressive loads can induce high local me- chanical loads in active material particles, and ultimately be detri- mental to cell performance. For example, Gnanaraj and co-workers 14 studied the effect of compression of anodes com- prised of KS-6 graphite and 10 wt % PVDF binderand LiCoO 2 cathodes, using voltammetry, electrochemical impedance spectros- copy EIS, and ex situ atomic force microscopy AFM. Electrodes were compressed at 5000 kg/cm 2 , using a rolling machine or hy- draulic press. The specific capacities were obtained from studies of cyclic voltammograms, in the order of highest to lowest, in un- pressed, hydraulically pressed, and rolled electrodes. Higher ac im- pedance was associated with low kinetic response. Reduced inter- particle distance may inhibit Li-ion diffusion among particles and electrolyte and lead to higher ac impedance. Thus, the efforts of Maleki et al. 13 and Gnanaraj et al. 14 suggest that there may be op- timal levels of compression, balancing conductive and mechanical properties. The structure of carbon matrices in the anode also strongly af- fects electrochemical performance. 4 Both highly disordered hard carbons and highly ordered graphitized carbons have been shown to produce high capacity. But they both suffer some degree of irrevers- ible capacity loss ICL, which is putatively related to the surface structure of carbonaceous materials and electrolyte systems. The most commonly used high-permittivity electrolyte solvents include propylene carbonate PCand ethylene carbonate EC. Each has inherent difficulties. EC-based electrolytes are believed to inhibit exfoliation in Li-ion cells with well-crystallized graphitic anodes; however, this phenomenon has not been observed for all types of graphite. 4 One well-known disadvantage of EC electrolytes is that they cannot be used at operating temperatures below -20°C, due to their relatively high freezing point 38°C 15,16 and rapid dropoff in ionic conductivity at low temperatures. 16 Ternary or even quanternary mixtures of solvents, such as mix- ture of EC, dimethyl carbonate DMC, and PC systems, are com- monly used to circumvent these problems. 15,16 However, decompo- sition of PC-based electrolytes on graphitic surfaces poses a major impediment to their usage. PC reduction produces severe gas forma- tion, creating high, localized particle surface pressures. Solvent per- colation in the resulting particle surfaces results in insulation of large portions of the active mass. 17 Thus, suppression of decompo- sition of PC-based electrolytes is highly desirable. Recently, core- shell types of graphite 18-20 have been used to exploit the turbostratic structure of the carbon coating as illustrated in Fig. 1 to reduce PC cointercalation of Li-ion, and to reduce ICL. * Electrochemical Society Active Member. z E-mail: sastry@umich.edu Journal of The Electrochemical Society, 151 9A1489-A1498 2004 0013-4651/2004/1519/A1489/10/$7.00 © The Electrochemical Society, Inc. A1489