Parameterisation of OCV and Diffusion Coefficient

Authors: Marcus Tuchel, Prof. Emma Kendrick

School of Metallurgy and Materials, University of Birmingham, Elms Rd, Birmingham B15 2SE

Abstract

To simulate a battery, the open circuit voltage (OCV) and diffusion coefficient of its active materials must be determined.
The established methodology is the Galvanostatic Intermittent Titration Technique (GITT) [1]. However, its significant limitation is the large experimental duration, requiring multiple days.
Here, Intermittent Current Interruption (ICI) was used as a fast alternative [2]. The rest durations were varied to optimise between accuracy and test duration.
OCV was found by extrapolating the voltage responses during the rests to their equilibrium values.
Subsequently, the extrapolated OCV was used to calculate the solid-state diffusion coefficient.
Testing indicated that OCV with an acceptable error could be determined using 10- minute rests, with an 84% time saving over GITT.
The diffusion coefficient recorded with ICI corresponded well with GITT but at marginally higher magnitudes.
These findings indicate ICI can be used to find the OCV, but the diffusion coefficient may need tuning if it is to be used in physics-based models.

Methodology

Figure 1.: Plot showing a linear regression fit to an
ICI rest which is extrapolated to find the OCV.

Materials: 2032-type coin cells, Li metal anode, LiFUN NMC920503 cathode, Celgard h1609 separator, 1M LiPF6 in EC:EMC 3:7 v/v with 2% VC electrolyte.
GITT: 10-minute C/10 pulses and 3-hour rests.
ICI: 5-minute C/10 pulses and rests varied between 5 s and 1 hour.
Linear regressions were fitted to voltage responses during rests against square root time after 1 s, an established methodology, Fig. 1 [2].
OCV was found by extrapolating the linear regression to t = 3 hours, the length of a GITT rest step.
Diffusion coefficient was found by applying the conventional GITT methodology but with the extrapolated OCV values from ICI.

Extrapolated OCV

GITT showed the best agreement with the pseudo-OCV with an average root mean square error (RMSE) of 7.36 mV, shown in Fig.2.
Shorter rest times led to worse OCV predictions, Fig.5.
The ICI test with 10-minute rests gave an acceptable RMSE of 13.50 mV in 16% of the time for GITT with double the number of data points, Fig. 4.

*Figure 2.: GITT OCV and pseudo-OCV alongside the RMSEs between them.*

*Figure 3.: 2-min rest ICI OCV and pseudo-OCV alongside the RMSEs between them.*

*Figure 4.: 10-min rest ICI OCV and pseudo-OCV alongside the RMSEs between them.*

*Figure 5.: Average RMSE between extrapolated OCVs and pseudo-OCVs against current interruption rest duration.*

Li ion diffusion coefficient

All of the diffusion coefficients calculated from ICI show the same trend as the values determined with GITT.
Values acquired with ICI are at greater magnitudes than values from GITT.
All of the ICI values recorded using rests between 15 minutes and 1 hour correspond very closely.
As rest durations of ICI tests shorten, the spread of the diffusion coefficient increases, giving a worse signal-to-noise ratio.
ICI with a 15-minute rest gave a clear response in only 21% of the time required for GITT whilst gathering double the number of data points.

Figure 6.: Diffusion coefficient against voltage calculated during charge and discharge from GITT
and ICI with the rest durations listed.

Conclusions and Future Work

ICI extrapolation can be used as a fast parameterisation technique to determine the OCV of a material.
- This facilitates faster parameterisation of battery materials for physics based models.
- Additionally, this approach could be used as part of reference performance tests to monitor changes to cell balancing during ageing.
The diffusion coefficient calculated using ICI was highly comparable to GITT but recorded much faster. However, it was at a slightly higher magnitude.
- This method could be used as a parameterisation technique to quickly find the diffusion coefficient of materials for use in physics-based models.
- With the slight difference in magnitude, tuning would be necessary.
Subsequently, this technique will be applied to other materials.
- Using the findings, it is anticipated that an appropriate ICI rest duration could be reliably predicted with some prior material information.

References

W. Weppner and R. A. Huggins, Determination of the Kinetic Parameters of Mixed‐Conducting Electrodes and Application to the System Li₃Sb, 1977 J. Electrochem. Soc. 124 1569
Z. Geng, T. Thiringer and M. J. Lacey, Intermittent Current Interruption Method for Commercial Lithium-Ion Batteries Aging Characterization, in IEEE Transactions on Transportation Electrification, vol. 8, no. 2, pp. 2985-2995, June 2022

Research Bio

Marcus Tuchel is a PhD student in the Birmingham Energy Materials Group, under the supervision of Prof. Emma Kendrick and Prof. Gerard Fernando. He is an associated PhD as part of the Faraday Institution MSM project. His research focuses on investigating Li-ion batteries using in-operando sensors to measure processes directly and novel experimental techniques to explore mechanisms.

He hopes to use these methodologies to optimise the formation process, determine a battery’s state of health, and for the early detection of faults.

Note: This poster has been published here with the kind permission of the University of Birmingham. All images and content are copyright of the University of Birmingham and cannot be reproduced without permission.