How Much Lithium is in a Li-Ion Vehicle Battery?

A very simple question!?You’d think that finding an accurate answer to this question would be dead easy, with five spare minutes and Google…and you would in fact be dead wrong in that assumption!

Why do I care??Why should you care??Because we’re constantly seeing rubbish articles with titles like “Lithium- the New Gasoline” etc.?People, some of them interested parties (investors in lithium stocks, or people short-selling Tesla stock etc.) are obviously concerned greatly with the quantities of materials required to make batteries for battery electric (BEV), plug in hybrid (PHEV) and hydrogen fuelcell electric vehicles (FCEVs).?We do also see lithium ion batteries starting to be used in grid storage/grid balancing applications (an application I don’t think is a good match for Li-ion batteries, but that’ll be the subject of a future article).?It seems, the mere fact that the term “lithium ion battery” contains the word “lithium” is sufficient reason for many to jump to the conclusion that the primary concern for battery materials cost, availability or environmental impact would be lithium.?If we want to determine whether or not there are problems with lithium, the first question we have to answer is how much lithium we’ll need to build out an EV battery fleet.?

The first link that comes up in a Google search for “how much lithium is in a lithium ion battery” is invariably an article on www.batteryuniversity.com – an excellent site sponsored by a Canadian firm (Cadex Electronics), but not always the best with providing underlying references for what is posted there.??Best also to ignore all comments to the articles there, as they are not moderated and hence tend to be complete rubbish:

https://batteryuniversity.com/learn/article/availability_of_lithium

This source does indeed give some data:?it claims that a Nissan Leaf battery has about 4 kg of lithium in it.?Assuming the author is (or was) talking about the 24 kWh nominal capacity Leaf battery, that’s about 167 g of lithium (in the battery) per kWh of nominal capacity, which it turns out isn’t far off the nominal value for Li ion batteries when you dig further into the literature.

With a little searching, you will however find numerous other values.?You will, as I did, find other data in the literature and in on-line articles which leads to calculations and estimates of fairly widely differing values.

My first attempt to answer this question was simple and direct:?I first searched the websites of, then directly contacted Tesla, General Motors and Nissan Canada and asked them the question.?Here is the response I received:

https://www.youtube.com/watch?v=K8E_zMLCRNg

Obviously the OEMs of BEVs don’t care to answer this question!?Why not??Do they consider the information “confidential”??Or are they trying to hide something??My feeling is that they’re just overwhelmed with requests for information and don’t have the resources or staff to answer them.?Irrespective of the reasons, the fact that they’re not providing any answers is a shame in my opinion, as the misinformation out there is driving people to incorrect conclusions about the future of their technology.

OK, so again we have to roll up our sleeves and do some work- work that shouldn’t be necessary.

Theoretical Minimum Lithium Content

It’s easy to figure out the minimum amount of lithium necessary to deliver a certain amount of energy.?Lithium has an atomic weight of 6.94 g/mol, so we already know that a little (mass of) lithium goes a long way.?You get one electron per lithium atom, and there are 96485 coulombs per mole of electrons.?One ampere is one coulomb per second.?Rearrange in a line calculation and you end up with 3.87 Ah per g of lithium, which is the theoretical maximum.?That gives you between 70 g of Li per kWh for a 3.7V nominal Li-NMC or Li-NCA battery, or 80 g/kWh for a 3.2V nominal LiFePO4 battery.?These are obviously totally inaccurate values, since the utilization of lithium in any real battery can never be 100%.?However, you’ll notice that these figures are used in the IATA calculation required to estimate the amount of lithium in batteries for air shipment.

https://batteryuniversity.com/learn/article/bu_704a_shipping_lithium_based_batteries_by_air

The underlying estimate is 0.3 g Li per Ah, which for a single cell at 3.7 volts nominal works out to about 81 g Li/kWh.?Clearly the IATA calculation is an underestimate as it is barely enough Li to satisfy the theoretical minimum.

Parts of a Lithium Ion Battery

I’ll let the reader find one of the million cartoons showing the parts of a lithium ion battery, but here’s a brief summary.?Each battery consists of:

Cathodes:?each cell has numerous cathodes, each of which consists of an aluminum foil current collector coated on both sides with a thin layer of powdered cathode material glued together with a small amount of PVDF binder.?The cathode material is a fine powder of a lithium-transition metal oxide such as LMO (Li2Mn2O4), Li-NCM (Li (Nix Coy Mnz)O2 where x+y+z =1) or Li-NCA (Li (Ni0.8 Co 0.15 Al0.05)O2) and may have coatings, conductivity enhancement fillers etc. added.?Note that the coefficients for the mixed oxides are molar ratios, not mass ratios- but since the atomic weights of Ni, Co and Mn are so close, that doesn’t really matter much for scoping-level calculations such as those in this paper.?Cathode active material may be 25-33% of the mass of the cell, based on the best references I could find.?The current collector foils make up another 13-16% of the cell mass.

Anodes:?an equal number of anodes, consisting of a copper foil current collector coated on both sides with powdered graphite or other graphitic carbon material, with or without silicon added, also held together by a binder.?Anode active material makes up 10-16% of the cell mass, with the current collectors adding another 18-27%, i.e. there’s more copper in these cells than active anode material.

Electrolyte:?a solution of a lithium salt such as LiPF6 in an organic solvent, typically a mixture of ethylene carbonate (EC), dimethycarbonate (DMC), ethylmethylcarbonate (EMC), vinylene carbonate and other special additives proprietary to the battery manufacturer.?The electrolyte tends to make up on the order of 10% of the cell mass

Separator:?a thin plastic membrane made of polyethylene and/or polypropylene which separates the anode from the cathode, provides a pore space occupied by electrolyte, and is designed to prevent lithium metal dendrites from growing and forming an anode/cathode short circuit

Casing:?many layers of anode, separator and cathode are stacked, the anode and cathode current collector foils are each welded to a terminal, and then the assembly is installed in a box or pouch, or rolled into a “jelly roll” and put in a cylindrical can.?The casing encloses the components, keep them dry, support the terminals, and provide a means of pressure venting should the cell become short circuited or overcharged.?The case and terminals again make up about 10% of the cell mass.

Kushnir et al referenced below is the source for the mass % of the various components.

Where Does the Extra Lithium Go?

We know we’re going to need more than the theoretical minimum 70-80 g Li/kWh, but how much more?

Reading this publication by William Tahil of Meridian Research (2010), you can get very worried about how much lithium you might actually need.?A shorter summary version of that longer article is available here:

https://evworld.com/article.cfm?storyid=1826

Tahil estimates that the Li content of a real-world Li ion vehicle battery would need to be on the order of 2-3 kg of technical grade lithium carbonate per kWh of PHEV battery, which amounts to between 375 and 563 g of Li metal/kWh, i.e. 5-8 times the theoretical minimum required.?Tahil lays out the reasons for this expectation, but they consist of several factors:?

1. Cathode capacity dependence on current rate:?higher current draws result in a lower effective capacity in a real battery, as shown in Ragone plots of cathode half cells.?Whereas the capacity of a particular cathode material may be X at 1/10C discharge rate, it may be only 1/4X at a 2C discharge rate- similarly for charging.?He speculates that this alone could represent a need for up to 4x the theoretical Li needed in cathode material for a PHEV battery
2. Irreversible capacity in the anode, i.e. Li tied up in the anode material that cannot be liberated during reversible charging/discharging.?He states that the anode may have 50-200 mAh/g tied up in irreversible capacity, which he converts to 1/6 to ? the lithium in the anode, but fails to give a reference for these figures.?He speculates that much of this is the amount of Li tied up in creating the solid-electrolyte interphase (SEI) layer, a layer of electrolyte which has been reduced by contact with Li metal in the anode.?This layer is permeable to Li+ ions but protects the anode’s Li from further exposure to electrolyte.
3. Ohmic, polarization and mass transfer energy losses in the cell, which result in the loss of chemical potential energy as heat in the battery
4. Extra installed capacity required to avoid high and low states of charge (SOC) which can result in capacity loss, plus the practical need to provide excess capacity to ensure that the battery has adequate capacity.?He estimates that practical EV batteries must be 25% larger, and hence have 25% more lithium in them, than might be predicted from the nominal capacity.
5. Refining losses for the cathode material purification and manufacturing steps, which he estimates may be up to 30% of the raw Li from the mine.

When all these efficiency factors are chained together, Tahil arrives at his estimate of 5-8x the theoretical Li content, or 375-560 g Li/kWh for a PHEV battery.

Problems With Tahil’s Estimates

While I am happy to stand corrected if someone actually comes forward with a disassembly, acid digestion and ICP analysis of a commercial EV battery demonstrating otherwise, here’s where I think Tahil’s estimate goes wrong:

1. His paper was written in 2010 and hasn’t been updated.?Things have changed- a lot- in the decade between when his references were written and the current Li ion state of the art
2. Modern cathode materials in modern cell configurations have Ragone plots much more favourable than those he used in his paper.?Capacity is much less sensitive to drain rate with modern materials
3. The PHEV battery is the worst case for current drain (C rate) on discharge, given that it is much smaller in kWh capacity than a BEV battery will need to be while its power demand is the same, i.e. it must still provide as many kW to the inverter driving the motor as the BEV.?The inherently larger BEV battery operates at a lower C rate during discharge by a significant margin as a result
4. The permanent anode capacity loss to the SEI layer is significantly overstated in his paper by my estimation based on a simple mass balance and other literature, given that the entire electrolyte represents only 10% of the mass of the cell.
5. It seems to me that Tahil double-counts some of the Li losses.?Some of them are inherent in the measured capacity of the anode and cathode materials themselves, and others are taken into account in the advertised nominal capacity of the battery in the vehicle itself

Accordingly, Tahil’s estimates seem, to me at least, to be significantly in excess of the real Li requirements for real BEV batteries. But Tahil does have a good point: you need to recover more Li from the earth than ends up in the battery itself. How much more? Regrettably, neither he nor I have a good reference for that figure, though his 70% efficiency figure is probably not too far off.

Dunn et al, Argonne National Laboratory (2012)

This paper by the ANL, authors of the GREET model frequently used for lifecycle analyses in the energy industry, shows up as a link in a ResearchGate page on the subject, again fairly high in the Google search results:?

https://greet.es.anl.gov/publication-lib-lca

Dunn et al’s paper gives some estimates related to a LMO battery for a BEV:?28 kWh, 463 lbs (210 kg), 33% of its mass being LMO cathode material- sounding suspiciously like a Gen 1 Nissan Leaf battery.?From this we can calculate a Li content of about 95 g Li/kWh.?However, if it is indeed a Gen 1 Nissan Leaf battery, the actual nominal capacity is closer to 21.3 kWh with a nameplate capacity of 24 kWh.?

https://en.wikipedia.org/wiki/Nissan_Leaf

Using 21.3 kWh, and adding another 10% for Li in the SEI and electrolyte as LiPF6, we end up with about 137 g Li in the battery per /kWh effective capacity.

Kushnir et al (2015)

Duncan Kushnir of Chalmers University in Sweden provides some estimates in this 2015 presentation linked here:

https://publications.lib.chalmers.se/records/fulltext/230991/local_230991.pdf

The recent Swedish study of lifecycle and embodied energy and emissions for EV batteries references Kushnir’s paper, which I dug up when doing the research for my own LinkedIn article about embodied energy in EV packs:

In particular, Kushnir provides this graph, which links to a previously released paper of his from 2012 which is behind Elsevier’s paywall so I’m not going to bother to pay to download it:

This estimate is about 160 g Li metal in the battery/kWh or roughly 2x the theoretical minimum Li content.?As you’ll see, this seems to line up with the rough magnitude of estimates from other sources.

It is interesting to note that in Appendix of Kushnir’s presentation, compositions of batteries are given for resource recovery through recycling.?Final cell capacities are given in Wh/kg for NMC333, LFP and LMO batteries, as well as the mass of cathode per unit mass of finished cells (which are roughly 25-27% for the cells of interest).?From the figures in his Appendix A, one calculates a range of 106 g Li/kWh for NMC333 to 137 g/kWh for LMO, plus about 10% for the Li in the salt in the electrolyte.?When one further crunches the numbers in his Appendix B, Li content available for recycling works out to 80-94 g Li in the battery/kWh in total including cathode and electrolyte- clearly lower than realistic, given how close these are to the theoretical minimum.

Goldman-Sachs Estimate

People are lazy and like to receive their information fast these days- even if that sacrifices accuracy...and unfortunately, there are people ready to satisfy that apparent need.???A number of rather slick-looking infographics on the topic of battery materials have been prepared by the Visual Capitalist, and these turn up fairly high in a Google search on this topic.?Unfortunately, these infographics aren’t even consistent with one another much less with other estimates, nor do they link their sources so they can be fact-checked.

Let’s take two examples:

This one, from 2016, is scattered all over the Internet.

https://www.visualcapitalist.com/critical-ingredients-fuel-battery-boom/

It contains a claim that “Goldman Sachs estimates that a Tesla Model S with a 70 kWh battery uses 63 kg of lithium carbonate equivalent (LCE)”, which appears to come from the following 2015 report (p. 17) with the sexy (and totally misleading) heading title, “What if I told you…Lithium was the New Gasoline”:

https://www.goldmansachs.com/our-thinking/pages/macroeconomic-insights-folder/what-if-i-told-you/report.pdf

The source of that Goldman Sachs estimate is anyone’s guess- they might tell you if you paid them.?But worked backwards, it amounts to about 209 g Li metal in the battery/kWh.

Then there’s this one from February 2017:

https://www.visualcapitalist.com/lithium-fuel-green-revolution/

This infographic includes the suggestion that a Tesla Model S contains “…up to 51 kg of lithium”.?Since the largest Model S battery is a P100D with 100 kWh nominal capacity, that would imply a use of about 510 g Li/kWh- all the way up at the high end of Tahil’s estimates.??Clearly, the authors here probably confused lithium METAL (Li) with lithium CARBONATE (Li2CO3), which along with lithium hydroxide is the primary material of commerce in the lithium industry- and which is only 18 wt% lithium.?Taking that into account, the estimate would be 510*0.18 = 92 g Li in the battery/kWh, which is probably light.?Perhaps this 51 kg of lithium was for a 70 kWh battery??Who knows- no references are given.

The popularity of pretty, easily-shared and frequently inaccurate information like this is very concerning to me.?It would not at all surprise me if public policy and investment decisions are being made daily on the basis of garbage like this- on rumour and speculation supported by confirmation bias, rather than on the basis of good measurements or data given by the people who know this best:?the OEMs actually making EV batteries.

So…How Much Lithium is in a Lithium Ion EV Battery?

The best estimate is around 160 g of Li metal in the battery per kWh of battery, or if you prefer, about 850 g of lithium carbonate equivalent (LCE) in the battery per kWh.?Again I am happy to stand corrected if someone actually comes forward with a disassembly, acid digestion and ICP analysis of a commercial EV battery demonstrating otherwise.?So far, the papers I’ve seen which do such analyses fail to provide the data necessary to compare the mass of Li to the original power storage capacity of the batteries. If you want, you can divide that by Tahil's estimate of 70% recovery in refining and turn that into 230 g of Li recovered from the earth per kWh of battery produced, though again that's a very sketchy figure.

By this measure, the E-Fire's 18.5 kWh LiFePO4 pack has about 3 kg of lithium in it as lithium. That lithium, as noted in my previous papers, eliminates the need to mine and refine many thousands of kg of crude oil into gasoline, and all the emissions associated with those activities plus the burning of that gasoline. And at the end of the battery's life, should lithium become scarce enough to make it valuable enough such that this is worthwhile, the lithium is all still there in the battery, available for recycling.

In my next article, I’ll have a look at the availability of the component metals in EV batteries, including the one which is scariest to anyone who, as I do, hopes for a rapid and broad adoption of battery EVs in the future.?That would not be lithium, but rather cobalt…

2020 year-end update: solid state batteries will ultimately use solid lithium metal anodes rather than graphite. This will mean that the Li use per kWh for lithium ion solid state batteries will be HIGHER- by an amount that will vary, but which can easily be 50%.

https://www.dhirubhai.net/pulse/part-2-battery-materials-we-dont-need-worry-paul-martin/