A Battery Taxonomy
Putting different battery chemistries in their place
Every single battery — from Allessandro Volta’s voltaic pile to whatever the latest startup is touting — requires three things: a cathode, an anode, and an electrolyte separating the two. What these components are made of is completely up to you.
We humans have come up with a lot of materials to help store energy in electrochemical devices: from ceramics in solid-state batteries to liquid-electrolytes in flow cells, to oxygen-cathodes used for metal-air — the list goes on. The materials you choose for each component, and the way you arrange them geometrically,1 will dictate your battery’s capabilities.
It would seem we could classify batteries based on their component states of matter to draw out some heuristic insights. This way, we can guestimate some general performance specs at a glance. (A TL;DR summary is at the bottom if you’re too lazy to read
Taking inspiration from the “The Cube Rule” for food, we present our Battery Taxonomy.
Instead of identifying food based on where the solid carb was, shall we try to come up with an equivalent for batteries based on whether the electrodes and electrolytes were solid, liquid, or gas?
Bit of background first:
1) Most common taxonomies are based on their performance criteria. For rechargeable batteries, we care about how much energy it holds, how quickly we can use or return the energy, and how many times we can repeat that process without killing it. Of course, this is all subject to the cost and danger levels tolerable by the end-user.
“In a lithium-ion battery, one in thirteen atoms gets up and out of its place once a day and moves 100 microns to the other side. That physical dislocation, back and forth, can cause massive damage on an atomic level.”
This picture also ties well with Prof. Dan Steingart’s “Stat Mech” Treatment of Batteries, which I seriously encourage you to read if you haven’t already. Instead of atoms, he considers the bonds between them:
A battery can have energy bonds or structural bonds: energy density depends on their ratio
Power density is determined by how quickly bonds can be formed/broken
The more the structural bonds are transformed, the harder it is to put the bonds back – worsening lifetime
Let’s try our best to mash these concepts together into our state-of-matter battery taxonomy, starting with the electrode, then the electrolyte, before running through some examples.
The following statements are not irrefutable laws of nature, so expect inadvertent exceptions that prove these rules.
⚡️ Electrodes: the energy storage medium
The only strict requirement of an electrode is that it conducts electrons into and out of an electrochemical system.
Electrodes are considered inert if they exclusively do that, or reactive if they themselves are converted in redox reactions. Inert electrodes can either have or not have an electrochemical fuel that releases/stores energy through reactions. Finally, these electrodes can be in a solid, liquid, or gas form.
Sketching this out:
A battery can have separate categories for each half cell. For example, zinc-air cells have a reactive-solid anode (metallic Zn) and an inert-gas cathode (porous carbon + O2).
Familiar lithium-ion systems generally fall into the reactive-solid category, with most electrode materials being some kind of solid.
Energy density improves when you improve your energy-to-structural bond ratio. Intercalation compounds, where Li+ slots into a host structure like graphite and NMC, have a lower bond ratio than conversion compounds like silicon, and of course, for Li-metal electrodes, every single bond is an energetic bond.3
Cycle life follows the inverse. A higher degree of electrode transformation means that a higher level of structural or “morphology” management is required to prevent degradation. Silicon expands a lot on cycling. Even worse, lithium metal will grow into uncontrollable morphologies unless you constrain it — many folks are trying to do this with stiff solid electrolytes. Meanwhile, electrodes like LTO are pretty bullet-proof when it comes to cycle life as they barely transform during charge & discharge.
Reactive-liquid electrodes can exist under the right conditions of temperature and pressure, but I can’t really think of any reactive-gas examples.4
Inert electrodes, as the name suggests, are very durable like carbon felt in vanadium-flow batteries.
Oxidant/reductant fuels store the energy for inert electrodes. Flow batteries use liquid electrolytes, while hydrogen fuel cells use gases. Inert-with-solid-fuel electrodes could also be thought of as reactive electrodes, like LFP (active fuel) that relies on a layer of electron-conducting carbon powder (inert electrode) to function.5
Flow battery electrolytes have an energy-to-structural bond relationship determined by solubility. Energy density is given by how much active material you can dissolve into the solvent, which is much lower than what you can pack into a crystal structure.6 Conversely, solvation also means the structure can be easily rearranged: no real atomic electrode damage occurs from physical dislocation.
Electrode reactions that involve gas, liquid, and solid together are slow and need expensive catalysts.7 These reactions must not only occur at a triple-phase boundary but also at a catalyst site like platinum coatings for fuel cells and cobalt oxides for lithium-air batteries. Of course, gas is as light as it gets, but what’s not so light is the equipment necessary to separate O2: you have to make sure the other 79% of air isn’t going to cause clogging side reactions.
Inert-no-fuel electrodes aren’t batteries, rather they are supercapacitors, like graphene cells that only store surface charges.
🍋 Electrolytes: the ion superhighway
Electrolytes can be solid, polymeric, liquid, or some mixed matrix of all states.8 All electrolytes need to allow ions to move between the two electrodes while preventing electrical contact between them.
Ion transport characteristics are arguably more sensitive to electrolyte geometric arrangement within a cell. While Figure 4 lists some properties associated with the state of matter, even though ions move faster through liquid, you may be better off with a thinner ceramic if your electrodes short circuit from being too close in a liquid.
Polymeric exchange membranes can be ion-selective so that only specific ions can pass through to react, as seen in fuel cells and some flow batteries.
Of course, safety gains can be had with careful electrolyte selection. This gets more important as we try to cram more energy and power into our batteries.
Gas in your electrolyte is usually a big problem. Are there any ion-conducting gases that can be used in a battery?9
Alrighty, let’s see if we can use this to work through some example battery types and draw out any heuristic insights into their characteristics and use cases…
🔋 Lithium batteries
Li metal is the most energy-dense anode: [reactive][solid]
Every single bond is an energy bond, meaning you can get a lightweight battery that could power electric planes. However, uneven lithium stripping and plating severely affect the cycle lifetime, so companies have looked at pairing the electrode with special ionic liquid or ceramic structured electrolytes to manage the lithium morphology.
LTO is a low energy anode with zero-strain conversion: [reactive][solid]
Sure the energy is low, but it can cycle at high rates for as long as you want because there is nearly no transformation in its structure. When the total cost of lifetime operation matters, LTO thriving for 12,000+ cycles could see it being used as a grid-storage option.
Li-Sulfur cathodes go through complex transformations: [reactive][solid]
Sulfur doesn’t just react with lithium, it goes through a series of polysulphides from S8→S6→S4→S2. Shuttle degradation occurs in these cells because you can lose part of the structure at every step. That said, the cost and theoretical energy offered by sulfur remain attractive, and some have looked into “caging strategies” to manage the sulfur morphology.
🧂 Hybrid-flow “metal-halide” batteries
Commercialized batteries of this sort usually come in hybrid-flow formats: like zinc-bromide or iron-chloride chemistries, though non-flow formats also exist. The following points see metal-halide batteries deployed in stationary storage scenarios:
Metallic anode [reactive][solid]
A more energy-dense solid phase, but dendrite growth can cause internal shorts, so many of these cells require reconditioning with deep discharges every few cycles to fully dissolve the zinc or other metallic anode to prevent degradation. Solid anodes also trade-off power capability with morphological uniformity.
Liquid electrolyte cathode [inert][redox fuel][liquid]
Energy density is lower as you can only get so much salt or coordination complex into the solution. But, the liquid phase means no morphology constraints and you can just store the liquid separate to where it reacts at the inert electrode.
🌬 Metal-air batteries
Comes in both flow and no-flow flavours, which could dictate whether it powers a single hearing aid or an entire neighbourhood. Many commercially relevant air-chemistries from lithium to zinc to iron to aluminium.
Metallic anode [reactive][solid]
Very similar characteristics to the above metallic anode sections10
Air cathode [inert][redox fuel][gas]
You need fancy catalysts to overcome the slow kinetics, high overpotentials, and low efficiencies of reactions simultaneously between solid, liquid, and gas. Also, many metal-air batteries will operate in a three-electrode setup, with two types of air-electrodes, because the fundamental reaction is different on charge and discharge: one for oxygen evolution on charge, and one for gas diffusion on discharge. If you’re using straight air, you also have to make sure you scrub out any impurities so that only oxygen is participating in reactions. Of course, it may all be worth it if your active material is literally a pile of rust that doesn’t need a catalyst to react at C/100…
🚰 Redox flow batteries
These batteries can be designed particularly well for larger and cheaper grid-storage applications. Vanadium is the current market incumbent with emerging chemistries such as metal-free aqueous organic. The flow-battery electrolyte design space is vast.
Liquid electrolyte electrode fuels [inert][redox fuel][liquid]
While energy density is limited by solubility, this “purist” form with liquid on each side of the battery allows one to fully decouple energy and power scaling. The redox-active electrolyte can be stored in tanks (energy) separate from the inert electrodes (power) themselves. Energy stored in solution means morphology is a non-issue and so these batteries can be cycled with deep discharges without damage.
🤯 and more…
👀 TL;DR summary
Electrolytes in solid or liquid form are also paired up with electrodes based on compatibility requirements.
Oftentimes, the phase of each material helps us understand the energy, power, and balance-of-plant characteristics of each battery half-cell electrode.
Researchers have been very creative in pairing up different phase-of-matter electrochemical materials to store energy:11
Companies have also been very creative in pairing up different phase-of-matter electrochemical materials to store energy:12
🌞 Thanks for reading!
📧 For tips, feedback, or inquiries, please reach out.
The way certain electrodes and electrolytes are arranged, along with their ratios, and whether they flow or not, can often do more to alter battery capabilities than the intrinsic material property themselves.
This transformation with atoms by breaking and making bonds is comparably harder to manage, for example vs. shuttling “massless” photons or electrons in solar-cell semiconductors.
Let me know if you do though, are there examples of redox-active and electron-conducting gaseous electrodes?
Carbon coated LFP that was able to reach its theoretical capacity was first presented at the 1999 ECS meeting by Ravet, Goodenough, Besner, Simoneau, Hovington, and Armand.
Electrolytes can generally get concentrated up to the 0.5 – 5 mol/L range, while Li+ concentration in graphite anodes easily reach 50 mol/L.
Kinetics get harder when you need simultaneous collisions between reactants and catalysts at a particular triple phase boundary.
Commercial LIBs use liquids soaked into a porous polymer separator (LP30 + Celgard), and some dendrite-blocking solid electrolytes blend polymer and ceramic together (PEO + LLZO)
No, I’m not counting liquified gas electrolytes, because they are liquified.