Lithium-ion batteries are used in countless applications today, from powering smartphones to electric cars. The main reasons for their popularity are their high energy density and efficiency, as well as a long-lasting battery life.  

Through R&D efforts, lithium-ion batteries continue to improve and optimize their safety, durability, energy storage, charging speed, and cost. However, this is often a complex and time-consuming process that must consider a large scope of components and variables that, combined, affect the performance of the battery.

Why is electrolyte viscosity important?

One key component to optimize when it comes to battery performance is the electrolyte. This is a solution that contains the ions that move from one electrode to the other, generating an electric current and releasing the energy stored on the battery.  

The properties of the electrolyte are essential to the performance of the battery. One of these properties is electrolyte viscosity.  

In general, a lower viscosity of the electrolyte is desirable to ensure better ion mobility and efficient battery performance. The reason is that a higher viscosity hinders the movement of ions through the electrolyte, which lowers the performance of the battery.  

When testing possible compositions for a new battery prototype, studying the viscosity of the electrolyte formulation is a crucial step to predict the performance of the battery.  

However, accurately measuring low viscosity liquids using experimental tools such as viscometers and rheometers can be quite challenging. In addition, taking experimental measurements for every single iteration of possible electrolyte formulations becomes extremely time consuming, especially when screening through hundreds or even thousands of potential compositions.

To overcome this challenge, full-atom molecular dynamics simulations can be used to predict the viscosity of any number of electrolyte formulations under the desired conditions.

Use case: Simulating electrolyte viscosity with QuantistryLab

With just a few clicks, the QuantistryLab platform allows the user to run molecular dynamics simulations of chemical systems, such as lubricants, polymers, and electrolytes, to model and predict many of their properties, including viscosity.  

To predict the viscosity of an electrolyte formulation, molecular dynamics simulations are deployed to measure fluctuations in the stress tensor over time. Using the Green-Kubo approach, a statistical mechanics method which employs autocorrelation functions of the stress tensor, an estimation of the viscosity at equilibrium is obtained. For very high viscosities, the platform ensures that the long-time behavior of the fluctuations is well captured by fitting a tail to the autocorrelation function.

Doing this is simple with QuantistryLab thanks to its Click&Simulate technology.  

The first step is to model the electrolyte. For this use case, a known electrolyte formulation was chosen, consisting of the following three compounds:

• Ethylene carbonate (EC)  

• Dimethyl carbonate (DMC)  

• Lithium hexafluorophosphate (LiPF6)

Using QuantistryLab, generating a model of a liquid, surface, alloy, and, for this use case, an electrolyte formulation, is straightforward.  The user can simply select the three components from the compound library and set the desired ratio.  

For this use case, the ratio was set as: 16 EC (C₃H₄O₃) + 270 DMC (C₃H₆O₃) + 77 LiPF₆  

Once the model of the electrolyte formulation is ready, its viscosity can be simulated with just a few clicks. The user can easily start a workflow run by selecting the viscosity feature, set the desired temperature and pressure, and start simulating.

The results yield a 3D atomistic model representing the dynamical behavior of all the atoms and molecules that make up the electrolyte formulation as a function of time.

The viscosity of the above electrolyte formulation, as predicted by QuantistryLab, is 5.47 mPa·s at 20 °C.

To validate the accuracy of this prediction, the model's viscosity value was compared to experimental results reported in a peer-reviewed publication. The experimental viscosity of the same electrolyte formulation at the same temperature and pressure was 5.17 ± 0.65 mPa·s.

The predicted value falls within the range of the experimental measurements reported in the literature. This result confirms the predictive power of the computational setup implemented by QuantistryLab.  

This use case demonstrates how to gather one data point of an electrolyte’s viscosity. By employing the “bulk simulations” option within the QuantistryLab platform, the user can screen a large set of systems at different conditions, thereby performing a full study of the viscosity of the electrolyte formulation of interest.  

These are some of the applications that this approach enables:

• Predicting the viscosity of an electrolyte that is challenging or costly to test in a laboratory setting

• Exploring the viscosity of an electrolyte under different conditions of pressure, temperature, and composition

• Investigating the performance of different ratios of an electrolyte’s components to find the optimal formulation for a desired application  

• High-throughput screening of electrolyte formulations to identify promising candidates before testing their performance in a laboratory setting

Simulations of viscosity can be extremely valuable when looking for candidates to formulate new electrolyte formulations for batteries. The key advantage of running simulations is the amount of time and resources that can be spared by predicting the electrolyte’s properties without performing hundreds of tests in a laboratory setting.

For this particular use case, looking at an electrolyte composed of EC, DMC and LiPF₆, a possible application would be to probe different ratios of these components to predict the potential performance of lithium-ion batteries with low amounts of EC in the electrolyte, which have been reported to perform exceptionally well at high voltage and in low temperature applications.  

‍