UHPC 101 Part 3 – Capacity Fade


The first two posts of the UHPC 101 series introduced Coulombic Efficiency (CE); the key metric used to compare and rank cells, and Coulombic Inefficiency per hour (CIE/hr); an extension of CE that allows comparison of cells cycled at both different rates and within different voltage intervals.  For NOVONIX, CE and CIE/hr are invaluable tools that enable making informed decisions on cell chemistries on a per-application basis. Since CIE/hr is obtained by normalizing CE for time, the underlying mechanisms that govern both are the same. It is then well worth it to dig deeper into the ingredients that make up CE! In this post, we describe the first of two components that make up the CE of a cell: capacity fade.

Capacity Fade

As batteries are cycled and age, the available lithium inventory irreversibly decreases. This process can be influenced by various mechanisms which themselves are affected by things like time, cycle number, temperature, operational voltage, chemistry, etc. The loss rate of accessible lithium is known as capacity fade.

Here is how we illustrate this concept:

Do you like maple syrup? Have you ever kept your maple syrup in a glass jar? If you have, surely you would have noticed some crystallization in the bottom of the jar – that is if the syrup lasted long enough!

Let’s zoom way out in time and consider what happens when the same jar is used as your daily maple syrup receptacle. Initially, the jar is full and there is no crystallization, but by the time the jar is empty, a small number of crystals have formed, and so you cannot get all the maple syrup out of the jar that was initially put in. You now refill the jar and realize that you cannot put quite as much in as the first time you filled it. And again, by the time the jar is empty, even more crystals have formed, and even more maple syrup was forever lost to crystallization (☹)


Initially, the jar is full and there is no crystallization, but by the time the jar is empty, a small number of crystals have formed, and so you cannot get all the maple syrup out of the jar that was initially put in. You now refill the jar and realize that you cannot put quite as much in as the first time you filled it.

The more the jar is emptied and refilled, and the longer the jar sits in the fridge, the more crystals form and the less space there is to store that delicious nectar. The reduction in jar volume available to be filled with maple syrup is analogous to the irreversible loss of capacity in a Li-ion cell.

So why is it that a Li-ion cell irreversibly loses capacity?

Electrolyte components can react at the surface of the negative electrode (usually graphite but can be Si or blend). For example, a solvent reduction reaction occurs when a Li-ion, solvent molecule, and electron react at the negative electrode surface. This reaction is irreversible and thus reduces the amount of available Li that can be inserted and removed from the electrodes during cycling, reducing the cell capacity.

This can be partially remedied with electrolyte additives that help reduce the rate of these reactions by improving the protective layer on the surface of the negative electrode. In addition to various electrochemical mechanisms, it is also possible for some of the electrode materials to become “inactive” due to particle cracking or loss of electrical contact from the current collectors which leads to “dead” material and even less available Li. Irreversible phase transformations of electrode materials can also “lock” lithium in place, further reducing the cell capacity.

Taken together, the loss of available Li for charge and discharge is referred to as Li “inventory” loss. The degree with which the discharge capacity of a cell diminishes over time indicates how much Li is irreversibly lost due to reactions with the electrolyte and active electrode material loss; this is called capacity fade. Determining the contribution of each mechanism towards the capacity fade can be elucidated with “dV/dQ analysis” with very precise measurements offered by UHPC.


The figure above demonstrates a real case study where differences in capacity fade for nearly identical cells are detected after only 25 cycles on UHPC tests. Two sets of pair cells are shown where the only difference between them is the choice of electrolyte additive; electrolyte 1 (black) contains no additives, while electrolyte 2 (red) contains 1 % of some additive. The discharge capacity versus cycle number (far-left panel) can be re-plotted in two ways that elucidate capacity fade: the middle panel shows the capacity loss each cycle, and the far-right panel shows the cumulative capacity loss as a function of cycle number.

After only 25 cycles, cells with 1 % additive (electrolyte 2), have lost ~0.001 Ah less than cells with electrolyte 1. Since both the positive and negative electrodes in these cells are the same, it can be deduced that electrolyte 2 leads to fewer irreversible reactions that consume Li (less “Li inventory loss”). It is important to note that all of these cells were cycled at the same rate. However, the concepts highlighted in our previous post (CIE/hr) can be applied analogously to account for cells cycled at different rates. Detecting differences in Li inventory loss between different cell chemistries in early cycle life is invaluable for quickly screening how electrolyte-electrode combinations influence Li inventory loss.

Capacity fade is only half of the CE story; charge end-point capacity slippage is the other half. A high-quality CE measurement can be decomposed into its two components, revealing underlying cell failure mechanisms. The next post will discuss charge end-point capacity slippage, a metric much less prevalent in common cell failure discussions, but nonetheless important. We will see that when combined, CE, CIE/hr, capacity fade, and charge end-point capacity slippage inform decisions at all stages of battery technology development programs, speeding up the whole development cycle – saving time, money and FRUSTRATION!

A future post will dive into the power of UHPC raw data analysis (beyond cycle metrics!).  With this detailed understanding of cell degradation, informed choices of electrode-electrolyte combinations can be made, reducing the size of trial-and-error testing matrices – effectively reducing the dimensionality of the cell material optimization problem.

Check our blog page regularly to keep up to date on the UHPC 101 series!

If you wish to suggest an article for this section, drop us a note below.
Let’s accelerate the battery supply chain revolution!