Lithium ion Batteries - in detail

Lithium ion batteries - How they work

Basic design and operation

Lithium is a very small atom – it is the third atom in the periodic table. Hydrogen is the smallest – one proton & one electron. Helium is the next: 2 protons in its nucleus and 2 electrons making a very stable outer “shell” of electrons, and as such it is unreactive chemically. Chemistry is all driven by what happens to the outer electrons of atoms – how stable their configuration is, how tightly bound they are to the nucleus of the atoms. On the smallest atoms the outer shell having two electrons is very stable, making helium essentially unreactive to anything.

Lithium has 3 protons in its nucleus and 3 electrons. Two of them form that beautifully stable 2-electron “shell”. The third is sort of spare. It can’t fit in the inner shell, and it whizzes around on its own, attracted electrically to the nucleus, but pushed further out from the nucleus by the inner two electrons. The next “level” out of electrons though, rather than taking two electrons, takes 8, so to be stable in terms of electron structure, lithium is quite keen to lose that outer electron and become a Li⁺ ion. There is no way, in electrical terms, that it could gain another 7 electrons.

Remember that hydrogen also has just one electron. Hydrogen is unusual in that it could in principle gain an electron to complete its outer shell and become an H^– anion, or it can lose its only electron and be just a free proton H⁺. However its single electron is bound more tightly in electrical terms to the nucleus than the outer electron of Lithium because it is closer into the nucleus.

Lithium metal is very reactive. If you drop some lithium metal into water it reacts, fizzing as it does so, losing an electron and pushing some of the hydrogen in the water molecules out, producing LiOH in solution and hydrogen bubbling off (which is why it fizzes). Essentially pure lithium is so reactive that it burns in water. Pure lithium metal will similarly react with any moisture in the atmosphere.

All lithium ion batteries make use of this desire of lithium to be rid of its outer electron and the relatively small size of the resulting Li⁺ ion. Clearly you do not want to have any water inside a lithium battery.

Surprisingly you can get lithium atoms to hide inside the structure of graphite. Carbon has four out of its complete outer shell of eight electrons, so is relatively happy to cope with spare electrons being around, but won’t actually react with an element trying to push an electron at it. It would far rather share electrons with neighbouring elements. A lithium atom with about six carbon atoms around it is a relatively stable setup, but the lithium is held fairly loosely in the structure. If the carbon is actually atomically in layers – ie graphite, the lithium can move through the carbon structure without disturbing it too much because the lithium is such a small atom. Within the graphite structure the lithium is still there basically as a lithium atom, and it really would be happier to lose its outer electron completely. Graphite, containing lithium, is therefore used as the Anode in a Lithium ion battery, rather than using pure lithium metal at the anode. This process of graphite absorbing the lithium is called “inculcation”. The lithium is inculcated into the graphite structure.

There are other chemicals which can hold lithium ions comfortably in their structures. These are usually oxides or phosphates of other metals, with structures which can either include lithium ions, or be a stable structure without the lithium being there at all. If the lithium is there then the lithium’s outer electron will be donated to one of the Oxygen or phosphate ions, but if the lithium is not there, the other metal will donate another outer electron to the oxygen or phosphate, leaving a spare electron to flow through the circuit to join back with the Li⁺ ion when it gets to the anode, when charging the battery.

Crystalline structures which do not change too much in size or shape are the ones which work in lithium ion batteries. The other metals are usually transition metals which have more complicated outer electron structures and have several different states, with different numbers of outer electrons, all of which are relatively stable.

To create an electrical join between the graphite containing lithium, and the oxide or phosphate which can cope happily with lithium ions, a liquid electrolyte is used. This is usually an organic solvent containing a lithium salt – something like LiPF₆. That means that the electrolyte will contain some Li⁺ ions in the solution which are relatively small and some PF₆^– which are bigger. The organic molecules are also fairly big. If you were to allow the graphite anode containing lithium and the other chemical at the cathode to touch, the force on the lithium in the graphite would be strong enough for the reaction to happen spontaneously – effectively creating an electrical short-circuit. Given that the lithium ions are small, a membrane which will allow only the lithium ions to pass through it slowly, is used to separate the graphite/lithium Anode from the metal oxide or phosphate Cathode. The thickness and porosity of the membrane and the concentration of Li⁺ ions in the electrolyte are designed such that no reaction happens spontaneously, so the battery does not discharge when it is disconnected.

When a wire connects the anode to the cathode, there is an easy route for any lithium in the graphite to lose an electron. A matching electron will then tend to be dumped into the metal oxide or phosphate crystal structure at the cathode end which will attract a Li+ ion out of the electrolyte solution and attract the Li+ ion just created in the graphite into the electrolyte solution.

So when the battery is creating charge, the graphite loses its inculcated lithium, which become Li⁺ ions, flow across the battery and get absorbed into the metal oxide crystalline structure.

When the battery is charging, electrons are pushed into the graphite and sucked out of the crystalline metal oxide or phosphate structure by the external charging voltage. Lithium ions in the electrolyte are electrically attracted into the graphite structure and the lithium ions in the crystal structure flow out into the electrolyte to replace them and can percolate through the membrane.

With an organic liquid electrolyte, you cannot use solid lithium as the cathode because tiny thread-like lithium “dendrytes” can form on the surface of the lithium which can pierce the separator membrane in the cell leading to an internal short and to thermal runaway.

The different designs of lithium ion batteries have different crystalline structures to hold the Li⁺ ions at the cathode end, use slightly different electrolyte chemicals, and might have different membrane structures and thicknesses.

Lithium ion batteries - generic problems

With an organic liquid electrolyte, you can’t use solid lithium as the cathode because tiny thread-like lithium “dendrytes” can form on the surface of the lithium which can pierce the separator membrane in the cell leading to an internal short and to thermal runaway.

If you charge a lithium ion battery, with a graphite anode, too quickly the surface/interface between the graphite and the electrolyte can become damaged as the lithium ions cannot be inculcated quickly enough. The surface thickens, and you can also end up with lithium plating on the surface. If that happens, then dendrites could follow.

The potential problems are – anything that damages the membrane between the anode and the cathode – if it tears or if any hole appears in it an internal short will happen and the battery will overheat and if the external casing is breached it will catch fire, causing other neighbouring cells to overheat and catch fire as well – this is a thermal run-away. The graphite is flammable. Any non-ionic lithium will react with any moisture in the atmosphere to produce hydrogen which will also burn, the organic solvents contain carbon and the membranes are usually made of some form of plastic. So once a Li⁺ ion battery catches fire, it is difficult to stop it. If it is fully charged, the problem is worse.

This was the problem that affected the first Boing 787 Dreamliners and caused a few fires, until Boeing redesigned the battery with active fan cooling.

The electrolytes used are volatile and combustible organic liquids which have a relatively narrow operating temperature. If the temperature drops below 0^oC the conductivity of the electrolyte reduces significantly and the efficiency of the battery reduces. If the temperature of the battery goes above about 60^oC the reactivity of the electrolyte can increase the rate of decomposition of other parts of the battery and lead to thermal runaway.

If the crystal structure or the graphite structure swells or changes shape too much when it has all the Lithium in it, or if the Lithium is pushed into, or pulled out of, the structure too quickly causing any deformation in the structure, then it can damage the physical structure of the battery and could damage the membrane – hence the advice never to over-charge or completely discharge a Li⁺ ion battery and the limitation on how quickly these batteries can be charged.

If the battery gets too hot the membrane can melt – this happens typically above about 90^oC at which point usually the problem is that the membrane stops letting any Li⁺ ions through. As a result the battery stops working all together, or massively reduces in efficiency. However if, as it melts, any hole were to appear in the membrane, then again the battery will short-circuit internally which could lead to thermal run-away.

Lithium ion battery types

The variations are usually due to different materials used for the cathode of the lithium ion battery. These change the energy storage capacity and typical voltage of the battery.

Lithium Iron Phosphate (LFP) 90 – 120 Wh/kg

Lithium Manganese Oxides (LMO) 100 – 150 Wh/kg

Lithium Cobalt Oxide (LCO) 150 – 200 Wh/kg

Li with Nickel Manganese and Cobalt oxides (NMC) 150 – 220 Wh/kg

Lithium with Nickel, Cobalt and Aluminium (NCA) 200 – 260 Wh/kg

It is also possible to use materials other than graphite at the anode

Lithium Titanate Oxide (LTO) 50 – 80 Wh/kg

Lithium Iron Phosphate (LFP)

LiFePO₄ cathode

The crystalline structure, both when the battery is charged – when there is very little lithium in it, and when discharged when there is as much lithium as iron in the structure, is virtually the same.

Within the crystal, for the PO₄, each phosphorus is surrounded by a tetrahedral structure of oxygens. The iron ions fit into octahedron arrangements of the oxygens, with one iron (Fe) to each phosphorus (P) in the overall pattern. The tetrahedron/octahedron structure, leaves spaces for Li⁺ ions to slip in along one axis of the crystalline structure. The iron changes from Fe²⁺ when the lithium is there as well, to Fe³⁺, losing electrons to the circuit, while the lithium is being pulled out of the structure towards the anode end of the battery as the battery charges.

LiFePO₄ occurs naturally – called triphylite, but for batteries there are many different ways to manufacture it in order to try to get the ideal crystal structure to allow the lithium to flow easily in and out of it and to allow the matching electrons to flow within the crystal as well.

The stability of the cathode crystal structure, in both its charged and uncharged state, means that these batteries are more tolerant to full charge and less prone to stress than other Li⁺ ion batteries even if held at a high voltage for a long time. However, they are heavier than other Li⁺ ion batteries, so not so good for Electric Vehicles.

Lithium Manganese Oxides (LMO)

Lithium Manganese Oxides can have several different crystalline structures. Used for the cathode.

LiMn₂O₄ has a 3-d grid like structure (called a spinel structure ) which is a bit like that of LiFePO₄. The oxygens in the crystal are basically hexagonally packed in layers. The manganese ions fit into some of the octahedral sites (3 oxygens above and 3 oxygens below. The lithium fits into some of the left over tetrahedral spaces, and can diffuse in and out of the grid structure without changing the overall structure very much, so the battery can charge and discharge, provided the crystalline structure is nicely uniform. The manganese copes with the lithium disappearing and reappearing in the crystal structure during charge/discharge by effectively changing its oxidation state swapping between being Mn³⁺ and Mn⁴⁺. At the surfaces of the electrode occasionally two neighbouring Mn³⁺ ions will swap to being one Mn⁴⁺ and one Mn²⁺ ion. However Mn²⁺ is soluble in most of the electrolytes used in lithium ion batteries, which means the cathode will degrade slowly over time.

However a lot of other lithium manganese oxides, with different proportions of lithium and manganese and less oxygen have layered structures, which include layers of Li⁺ ions, in a sequence O, Li, O, [Mn mixture] O, Li, O, [Mn mixture]. In these structures the lithium can flow more easily along the plain of lithium ions, but if too much lithium is pulled out of the structure, the structure could distort a bit. Each manganese ion is again surrounded by an octahedron of oxygens, 3 from each of the layers of oxygens on either side of the manganese.

In Li₂MnO₃ there are layers of pure lithium ions and layers of lithium and manganese mixed, in a ratio of 2:1, sandwiched between the oxygen layers. Batteries using this cathode can be charged to a higher voltage, but if this is done (if the battery is over-charged) the surface of the crystalline structure, can lose oxygen – which could ultimately lead to thermal runaway as the oxygen could react with the electrolyte.

LiMnO₂ is not useful as a cathode material on its own because it has a corrugated layer structure to the manganese layers which becomes unstable if it is charged and discharged.

Li₂MnO₂ has a similar layered structure to LiMnO₃. It is fairly stable when charged and discharged, and given its high proportion of lithium has a higher capacity to weight ratio. However it has the same problem of the manganese potentially dissolving from the surface because of the formation of Mn²⁺ ions.

Research is still ongoing into creating the best combination of lithium manganese oxides, trying to get multiple different crystal structures in the cathode, which has the benefit of the stability of the LiMnO₄ but the charge capacity of the Li₂MnO₂ and the rate of charge possible from LiMnO₃.

LMO batteries – faster charging, higher current discharging and increased thermal stability compared with Lithium Cobalt Oxide.

Lithium Cobalt Oxide (LCO)

LiCoO₂ cathode

Lithium Cobalt oxide, LiCoO₂ is again a layered structure with alternate layers of Li⁺ ions and cobalt ions sandwiched between layers of oxygen atoms. So the layers are oxygen, cobalt, oxygen, lithium, oxygen… etc, each layer in a hexagonal packed grid. The cobalt is nominally in a Co³⁺ state, but can be oxidized to Co⁴⁺ state fairly easily, so the Li⁺ ions included in the lithium layers of the structure can go down to almost zero, and electrons can flow out of the cathode to match the lost lithium ions.

The preference for the cobalt to fall back to being an Co³⁺ ion is a driving force to pull the Li⁺ ions back from the anode and to push the electrons through the circuit, when the charged battery is connected to a circuit. The structure at the cathode when the battery is charged is layers of cobalt with oxygen on either side of those cobalt layers, but far less lithium between those layers. However if the battery is overcharged the structure can distort at the atomic level, with the cobalt moving slightly out of its neat planar structure, which in turn will stop the lithium from moving in and out so easily. This means that the ability of the battery to charge and discharge deteriorates.

If the battery gets hot – more than about 130^oC, LiCoO₂ can decompose to produce oxygen which will react with the organic solvent of the battery to produce a yet more heat and expansion – the whole thing will go through a thermal runaway and catch fire. The tendency for it to decompose like this is greater if the battery is charged and there is less lithium in the cathode structure.

Drawbacks: relatively short life-cycle, less thermal stability, smaller load capabilities (ie they need frequent recharging)

Lithium Nickel Manganese Cobalt Oxides (NMC)

Li with Nickel Manganese Cobalt oxides (NMC) cathode

These batteries are the same structure as the LiCoO₂ batteries, but with some of the cobalt atoms replaced with nickel or manganese. These are all similar in terms of size and electrical properties to the each other.

The proportions of the three metals can be varied. Different manufacturers tend to use different proportions. The ions do not appear to form any regular pattern in the nickel/manganese/cobalt layers.

Nickel and manganese are both cheaper than cobalt. Nickel allows the battery to discharge more fully and allows electrons to move through the structure more easily, while manganese makes the structure more thermally stable and improves the battery life.

Ni Co and Mn can have states of Ni²⁺ or Ni³⁺, Co²⁺ Co³⁺, Mn³⁺ Mn⁴⁺

Depending on the proportions of nickel, cobalt and manganese used, the proportion of oxidation states of the metals in the structure to begin with, when the battery is discharged, varies. The larger the proportion of nickel that gets added into the mix, the more of that ends up as Ni³⁺ initially.

Although this makes the battery easier to charge and discharge, because electrons can flow through the structure more easily, it also makes the mixture less stable thermally and chemically. Another problem with too much nickel in the battery chemistry is that when making the crystal structure in the first place, occasionally Li⁺ ions and Ni²⁺ ions can end up in the wrong places, because although they are different chemically, from an ionic point of view as far as electrical attractions are concerned, they are a very similar “size”, so can swap places in the structure when it is being made. (LiNi disorder) If nickel ends up in the layer where lithium is supposed to be, it can stop the lithium ions flowing smoothly in and out of the crystal structure. Also if there is a lot of nickel in the mixture, then it is likely to become Ni³⁺ during the charging process, even if it starts out as Ni²⁺, so the battery will become more thermally and chemically unstable when it is charged.

In a well-behaved / well-ordered Li[NMC]Oxide structure, as the lithium ions get pulled out of their layers during the charging process, the remaining oxygen-metal-oxygen layers slightly repel each other when there are fewer Li⁺ ions between the layers to stick them together electrically. This means that the “gaps” through which the lithium has to pass open up slightly making it easier for them to move as the charging process progresses.

Co³⁺, Ni²⁺ and Mn⁴⁺ all have similar bond lengths with oxygen

Co²⁺, Ni³⁺ and Mn³⁺ distort the octahedral cells because the odd number of electrons in one of the outer orbitals makes the bond lengths to the oxygens different

Ni³⁺ & Co²⁺ both have a single electron in their e_g orbit. The e_g energy level of transition metals is very similar to the Oxygen 2p orbital so can form a covalent bond with the oxygen, whereas normally you expect the interaction between the metal ion and the oxygen to be an ionic bond. Clearly as the battery charges and discharges the cobalt, nickel and manganese will change their oxidation states, so the oxygens and NMC layers will try to distort very slightly when this happens.

Depending on the proportions of Ni, Co and Mn, NMC batteries can either have a high specific energy from the inclusion of the nickel, or a high specific power from mixtures with less nickel, but not both.

Lithium Nickel Cobalt and Aluminium (NCA)

LiNiCoAlO₂ cathode

These batteries have the same structure as the NMC batteries.

However, the aluminium will not change its oxidation state like manganese can, though the Al³⁺ ion is about the same size as the Mn⁴⁺ ion. In an NMC battery most of the Mn is in the ⁴⁺ state, so different proportions of the three metals will work for the mixture including aluminium.

These batteries are more expensive to produce, and less stable than other forms of lithium ion batteries, so they need active thermal and electrical monitoring to make sure they are working correctly.

They have high specific energy, good specific power and a reasonably long lifespan, in spite of their potential lack of stability. They are what Tesla usually use in their Electric Vehicles because of their higher energy density and fast charging capability.

Versions which have a larger proportion of nickel have higher energy density and can be charged more quickly. However, as the nickel content increases, the risk of degradation, structural instability and thermal breakdown increases.

If a battery is heated to 180^oC it will have a thermal runaway, however it if has been previously over-charged, the thermal runaway can happen as low as 65^oC. The aluminium ions in the battery increase the stability of the crystal structure, but reduce its capacity because they take no part in the oxidation/reduction process because their only oxidation state is Al³⁺.

LiNiO₂ which is a similar structure to NCA or NiO₂ cannot yet be used as a battery material itself because it is mechanically unstable, shows rapid loss of capacity and has even worse safety issues.

Lithium Titanate Oxide (LTO)

Li₂TiO₃ anode, in place of graphite

The advantage of Lithium Titanate for the anode material in a lithium ion battery, is that it has hardly any change in shape during the charge/discharge process. It is a spinel crystal structure, like LiMn₂O₄, and the lithium can move into the unoccupied tetrahedral gaps between the oxygens. The cathode of the battery could be any of the metal oxides used in other lithium ion batteries.

In an LTO battery, nano crystals of lithium titanate are used for an anode in place of graphite. These nano crsytals have a far greater surface area than the graphite, which make it possible to charge and discharge the batteries far more quickly – because the Li⁺ ions will be able to migrate into the crystal structures over the whole of that surface. Graphite has a surface area of about 3m²/g whereas the nano-crystals have a surface area of about 100m²/g. Being able to discharge quickly means they can supply relatively high currents.

There are no problems with surface thickening, or lithium plating even when charging quickly, which makes these batteries safer, and the anode will cope with a wider temperature range. Removing the graphite from the anode means there is less flammable material in the battery. These batteries do seem to be more stable than the ones with graphite anodes.

The disadvantages are that they have a lower inherent voltage (2.4V), compared with their graphite cousins (3.7V). As a result these batteries have a much lower specific energy. Their cost is higher as titanium is more expensive than graphite.