Thread by @realscientists

Right! It's time for me to settle in for the evening and talk about lithium batteries and where (I think) they're going. But I think a good start is to look at where they've come from.

So here comes a brief history of lithium batteries - accurate to the best of my knowledge!

Our story begins in the 1960s. The motivation for developing batteries was much the same then as it is now - pollution free energy storage, with portable devices and electric vehicles in mind. Lead-acid and nickel-cadmium were the gold standard for batteries - but too heavy.

Scientists were already starting to consider batteries based on Li or Na in the 60s, because theoretically they could give cell voltages of around 4 V instead of <2 V - but those tried either would only work at high temperature or had big problems with cycle lifetime.

In the mid-70s, Stan Whittingham (then at Exxon Mobil) showed how TiS2 could "intercalate" Li ions into its layered crystal structure (shown - from WebElements) - and how this could be used as the basis for a new battery.

Exxon looked at commercialising the TiS2 battery using a Li metal negative electrode, but unfortunately Li metal has a lot of problems - it makes a habit of growing filaments of Li metal through the separator, short-circuiting the battery and causing it to burst into flames...

Although it has some good properties (e.g. conductivity), TiS2 was not great for other reasons too - it only gave an average voltage of ~2.2 V, it's sensitive to air, and it really needs to be paired with a Li metal electrode, because it contains no Li in its structure.

All of these downsides were effectively solved in 1980, by John Goodenough (pictured; img cred: ECS) and his group, then at Oxford, with their work on lithium cobalt oxide (LiCoO2, or otherwise known as LCO).

LCO does have lithium in its structure, so no requirement to pair it with a Li metal electrode; its average voltage is much closer to 4 V, offering a real 'energy density' advantage over existing technologies; and it's air-stable, and still has acceptable conductivity.

A range of metal oxide electrodes were investigated around this time and LCO proved to be the best - another major one at the time that finds use still today is lithium manganese oxide (LiMn2O4, LMO) - some advantages/disadvantages compared to LCO, but I'll not go into this now..

A battery, of course, needs a negative electrode as well as a positive electrode - and since Li metal was ruled out, an alternative was needed.

The concept of the Li-ion battery, with "intercalation" materials used for both of the electrodes - so no Li metal - had been proposed much earlier. Intercalation compounds of graphite - with ions such as Li or K and many more besides were well-known in the 70s.

But for a long time, graphite was not considered usable - the Li ions had a habit of dragging solvent molecules from the electrolyte into the graphite layers, which would decompose into gases and destroy the layers.

Around the same time as Goodenough's lab made their LCO discovery, Rachid Yazami demonstrated reversible electrochemical Li intercalation into graphite. The voltage characteristics are great - graphite has very close to the potential of Li, and a rather good Li storage capacity.

However - Yazami had to use a polymer electrolyte, which only worked at higher temperature - still not ideal.

Meanwhile, Akira Yoshino's group at Asahi Kasei Corp in Japan started experimenting with various carbon and related materials paired with various metal oxide electrodes.

(IIRC now...) Yoshino's group found that lithium could be reversibly inserted and extracted from non-graphitic carbon, such as coke and so-called 'hard carbons' - and in 1991 Sony marketed the first Li-ion battery with LCO as the positive electrode and carbon as the negative.

Eventually, the problems with graphite destruction were solved with the discovery that using ethylene carbonate as one of the electrolyte solvents resulted in the formation of a film - the so-called solid-electrolyte interphase (SEI) on the graphite surface, which protects it.

The LCO-graphite Li-ion system became the standard for portable electronics, and it still is to this day. In terms of chemistry, commercial Li-ion batteries have not changed an awful lot since the 1990s...

What has changed considerably is the engineering - through consistent improvements in fabrication methods, Li-ion batteries almost tripled in energy density since the early 90s.

OK, it's not Moore's law, but the difficulty of this shouldn't be underestimated!

Now there are two major advances on the chemistry side which are starting to make their mark in EVs; the 'NMC' family of positive electrodes, and silicon as a negative electrode material...

NMC - Li(Ni,Mn,Co)O2 - is a family of compounds sharing the same layered structure as LCO but where Ni and Mn replaces some of the Co. These materials were investigated in the early 2000s and allow for higher capacities (and hence energy density) than LCO.

NMC compounds are usually referred to by a number - e.g. NMC111, NMC532, NMC811 - which indicates the ratio of the metal elements in the structure.

The Ni-rich versions (e.g. 811) offer the highest energy densities, but have the biggest problems with stability.

A relative of the NMC family is NCA (Ni,Co,Al) - an electrode material most notably used by Panasonic who supply a company called @Tesla, you may have heard of them...?

Meanwhile, at the negative electrode - silicon has long been known to alloy with huge amounts of Li - 10x as much as graphite - at around about the same electrode potential. Unfortunately, the alloying comes with huge particle volume changes which make for poor cycle life.

However, Si can be embedded into a composite with graphite - just a few percent of added Si can double the capacity of the whole negative electrode without harming cycle life. This is now finding its way into some commercial cells.

So to summarise this monster thread - we now have batteries which can store around 270 Wh/kg now, compared to ~80 Wh/kg in the early 90s. Back in the 60s, 220 Wh/kg was considered the target for electric vehicles to become a realistic proposition (guess they were right?!)

Of course, we're not happy with 270 Wh/kg now - we want more! Demands for 400, 500 Wh/kg are not uncommon now, but we're starting to feel like we're pushing at the upper limit of what Li-ion can achieve. Or are we? A lot of people have a lot of ideas for the future...