Author: Matthew Sullivan | Major: Mechanical Engineering
This semester, with funding from the University of Arkansas Honors College Research Grant, I – Matthew Sullivan – was able to continue my research with Dr. Xiangbo Meng and the Department of Mechanical Engineering with nickel-rich cathode material for lithium-ion batteries (LIBs). Through this research, I have developed a procedure by which lithium nickel manganese cobalt oxide (NMC, LiNixMnyCozO2) can realize higher cyclability due to an optimization of the electrode surface in LIBs.
At the start of this semester, the basics of printing using viscous NMC-rich slurries had been ironed out, and this semester was critical for further optimization of the process as well as data acquisition from printed and non-printed cells. NMC is highly sought-after as an electrode material for LIBs because it has a higher capacity than other traditional electrode materials such as lithium cobalt oxide (LCO, LiCoO2) – the most common cathode found in modern LIBs. Despite its high energy density and reduction in expensive chemicals such as cobalt in NMC when compared to LCO, NMC has been slow to reach the commercial sphere because of its instability at the microscale. The 2-dimensional layered structure in NMC is prone to collapse into a 3-dimensional spinel structure. This failure is commonly attributed to the cycling of lithium ions into and out of the cathode during discharge and charge of the battery. Furthermore, batteries where fast-charge is desired, such as those seen in electric vehicles, are even more susceptible to crystal failure due to the high areal current density. By 3D printing the electrode into a more complex 3D shape, the areal current density can be reduced with respect to traditional flat, planar electrodes.
3D Printed NMC 811 Sample
Figure 1: Printed Cathode Sample on Aluminum Current Collector
To test this hypothesis, two slurries were created with NMC 811 (80% nickel, 10% manganese, 10% cobalt) cathode powder, polyvinylidene fluoride (PVDF) polymer binder, and carbon black conductivity enhancer in a 90:5:5 weight ratio to create the solids mass. This solids mass was then mixed with an equal mass of n-methyl-2-pyrrolidone (NMP) solvent. This mixture was mixed at 2000 RPM for 60 minutes until a homogenous slurry was obtained. The slurry was then placed into the EMO-25 printing head on a Hyrel ESR 3D printer and printed with two different nozzle thicknesses. The resulting print is seen in Figure 1. On the right of the image shows the “thin-printed” sample which was printed through a 0.2 mm needle, and on the left shows the “thick-printed” sample which was printed through a 0.53 mm needle. It is important to note that the actual thicknesses are greater than the nozzle used due to the die swell of the material after stressing through the nozzle, but the coefficient of die swell was not able to be accurately determined with available technology.
The printed electrode was allowed to dry overnight and then transferred into a vacuum where it was heated and dried to ensure all NMP solution had evaporated out. Two coin cells were made from each of the printing thicknesses, and two coin cells were made from a traditional planar electrode with a thickness of 0.1mm for a total of six coin cells tested. The coin cells are currently undergoing testing at a low charging rate (C/5, where C = 200 mA/g) to determine the effects of printing thickness on the life span of the cells. The thin-printed samples have a mass-loading of electrode material that is too low to be properly evaluated, and so their data has been disregarded. Between the thick-printed samples and the traditional samples, the printed samples show higher capacity retention with an average 84% of its original capacity, compared with the traditional samples that have decreased to an 80% capacity retention. These values are still relatively similar, however, and so further cycling and testing is still required to confirm or refute the proposed theory. Currently, however, it is believed that by reducing the areal current density – the supplied energy divided by the electrode surface area – from 0.00152 mA mm-2 in the traditional sample to 0.00132 mA mm-2 in the printed sample, the layered crystal structure of the NMC has been better preserved in the printed samples.
The funding from the Honors College Research Grant has allowed for many important materials to be acquired for this project. Namely, the printing head used in this study, the Hyrel EMO-25, has been crucial due to the high dynamic viscosities realized by these cathode slurries. In addition, some of the funding went towards a new cycler for the lab that allowed for more accurate charge and discharge of the created batteries and new data acquisition methods used in the quantification of capacity and capacity retention. Further research into this project will see a shift towards anodic electrode material being tested in a similar fashion, with the hopes of developing and publishing the first paper on 3D-printing of NMC 811 cathode material and NMC 811 full cells.