Here Are The Battery Makers That Used The Most Cobalt In 2018

FEB 9 2019 BY MARK KANE 45

LG Chem used about the same amount of cobalt as Panasonic, but delivered less GWh

Cobalt is an important element for lithium-ion batteries used in electric cars. Manufacturers are developing new cells with less cobalt content (below 10% of cathode weight) to decrease costs and reliance on cobalt sources.

According to Adamas Intelligence, three top cell suppliers by total capacity in 2018 – Panasonic, LG Chem, CATL – deployed more cobalt than all the others combined.

The interesting teaser of the report reveals the Top five suppliers:

There is no BYD, as we believe it uses mostly LiFePO4 cathode material batteries.

We don’t have data yet for EV battery sales in 2018, but taking into consideration that Panasonic is way above the competitors (see results in 2017), we can draw some assumptions.

For example, LG Chem, which yet needs to switch to NCM 811 batteries (nickel:cobalt:manganese at a ratio of 8:1:1 0 10% of cobalt) is probably still using – on average – a relatively high portion of cobalt. On the other hand, it seems that Panasonic is using the lowest amount of cobalt per kWh of capacity. Low usage of cobalt in Tesla/Panasonic cells was hinted at by Elon Musk by the way of Q1 2018 report, which of course caught our attention.

The lower the amount of cobalt per kWh of cell, the lower – at least in theory – should be the material cost of the lithium-ion cell. However, specific properties of cobalt make it difficult to replace with other chemistry, so there could be a drawback in performance. Ongoing developments are on a new concept that would enable eliminating cobalt, without sacrificing performances of the cells.

Here are battery production results in 2017 – as all suppliers are expanding, the proportion between Panasonic and LG chem probably will not change significantly (Panasonic is still expected #1):

MWh of EV li-ion batteries produced in 2017 (data source: EV Sales Blog)

Source: Adamas Intelligence via Green Car Congress

Categories: Battery Tech

Tags: , , , ,

Leave a Reply

45 Comments on "Here Are The Battery Makers That Used The Most Cobalt In 2018"

newest oldest most voted

I frankly don’t believe the cells in Tesla cars have less cobalt than NMC 811. A 2017 ECS paper analyzed cell from a Model S and found they were standard 80/15/5 NCA (15% cobalt). No silicon in the anode, either. Tesla may have switched to a 83/13/4 cathode since then and/or started using anodes with small amounts of silicon, I’ve not seen a cell teardown since the ECS paper.

BYD now makes NMC alongside LFP. I’m surprised they don’t appear on this list.

SNE said CATL GWh/month output exceeded Panasonic’s in early-mid 2018. Panasonic moved back into the lead when Model 3 production finally ramped up at the end of June. CATL will make another run this year as Model 3 production turns flattish. CATL cobalt consumption is probably much higher than the graph shows, although they have NMC811 I think most of their output uses more cobalt.

> I frankly don’t believe the cells in Tesla cars have less cobalt than NMC 811.
> A 2017 ECS paper analyzed cell from a Model S and found they were
> standard 80/15/5 NCA (15% cobalt)

You’re aware that the batteries in Model S/X are entirely different from the batteries in Model 3, right?

Model S/X batteries are sourced from Panasonic’s existing output in Japan, and as far as I know, are not changed specifically for Tesla at all. They are indeed, as your paper found, completely standard cells.

Model 3 batteries are made by Panasonic, but in Tesla’s factory and to Tesla’s specification. Those are the ones that are claimed to be low in Cobalt. If the paper you read didn’t look at a Model 3 battery then they were wasting their time.

There is zero evidence they use a different chemistry in 2170s vs. 18650s. Everyone squawking about 30% improvement just embarrassed themselves. The specs are virtually identical once you adjust for geometry. Also, Tesla was making claims about chemistry (silicon, etc.) long before the Model 3.

There’s also absolutely no reason to use different chemistries. Why would they buy 8 GWh/year of an inferior chemistry? That’d be completely brain dead.

Tesla has already started reducing cobalt usage in 2016 models of Model S and X and further reduced it in Model 3.

Also, NCA 80/15/5 is not 15% cobalt. It is 9.2% cobalt.

Where did you get the 9.2% figure?…

I found it on a website, but looking at it in more detail it is a bit more complex.

Looking for example in this article that looks at panasonic 18650:

It shows 7.77% though it uses 82/15/3.

The thing is that it is based on composition, not weight, hence why even if it shows 15%, the amount by weight is less.

As far as I can tell, the weight of nickel and cobalt is pretty much the same (while aluminium is much lighter) — so I don’t see how cobalt should amount to less in weight…

The major differences I have seen in the wild are because some are comparing only the active cathode metals (which is what the composition refers to), while others are comparing the entire cathode mix (including oxygen), or more commonly, all active materials in the battery. In the latter case, the cathode metals make up only about 40% of weight, which is well in line with most figures I have seen. (The Model 3 cells for example have been determined to have ~2.9% cobalt contents, which works out to about 8% of cathode metals.)

Why don’t you actually look at 2170 cells tear-down instead of spewing BS.
Here, take this link.

I’ve seen this. I don’t think you understand what he said. He did not analyze cell chemistry. And the measurements he did perform indicate the 2170s Tesla uses have the same chemistry as their 18650s, exactly as I said. The Model 3 pack design is much different, but not the cell chemistry.

Also, Jack’s great but he REALLY needs a technical editor:
48:30 – “Model S is 660 cubic mm. The Model 3 is 970. That’s a 60% increase.”

He’s off by 25x on volume and the increase is 47%, not 60. Just one trivial example of the type of mistakes he makes when free-associating. He needs a good technical editor.

There is, in fact, solid evidence that the Model 3 cells use less cobalt. Even ignoring the fact that you are accusing Tesla of blatantly lying — not just putting a spin on things (as they are admittedly prone to doing), but rather straight-out lying — there was a teardown last year that confirmed <3 % cobalt contents in the whole cell, translating to ~8% in the cathode.

I’m not aware of this teardown and can’t find it via Google. Do you have links?

18650 cells with typical 80/15/5 NCA formulation are ~3.6% cobalt (0.14kg cobalt per kWh of battery). I’ve read Tesla went to a 83/13/4 formulation, which is ~3.1% cobalt. NMC811 is ~2.5% cobalt.

I’ve never seen Tesla claim 2170 chemistry is different than 18650.

Don’t have saved links, and don’t know how to dig them up easily…

The percentages you cite are *way* lower than what I have seen elsewhere. They don’t sound right to me, even if they include passive parts of the battery (casing etc.)…

Tesla did explicitly say on at least one occasion that the 21700 cells have different chemistry. They also explicitly claimed more recently that their current cells already have a lower cobalt contents than the upcoming generation of NMC cells. (Presumably NMC811.)

“There is zero evidence they use a different chemistry in 2170s vs. 18650s.”

a plain lie


zero evidence? The very fact that Panasonic produces almost 50% of the world’s batteries and yet, has cobalt usage similar to a company that produces less than 1/3 of the battery output pretty much says that you are way wrong.

That has no bearing on the difference between Panasonic’s 18650 and 2170.

The amount of 18650 sold over the last couple of years has been fairly constant. Why? Because Tesla has sold around 50K MS/MX.
OTOH, the 2170’s which uses Tesla’s formula and less than 1/2 of the cobalt that 18650 does, was where all the growth was over the last couple of years. So yeah. Related.

I don’t know if Panasonic is using a different chemistry for Tesla’s cells, but certainly they aren’t standard off-the-shelf cells. Tesla hasn’t used off-the-shelf cells since the original Roadster.

If nothing else, the 18650 cells Panasonic makes for Tesla lack certain safety features that are present in consumer-grade cells… and I would guess the same for the 2170 cells. Tesla also has patents for the internal structure of cells. I don’t know if Panasonic is using those patents or not, but certainly a different internal arrangement isn’t going to show up on somebody’s chemical analysis of a cell.

This article makes a lot of assumptions which it treats as facts… but then, so do the different comments in this sub-thread. I submit we need a lot more skepticism and questioning of assumptions/premises, and a lot less dogmatic assertion of opinions as fact.

AFAIK CATL still makes a lot of LFP cells (at least for buses — not sure about passenger vehicles); so it’s to be expected that their cobalt use is relatively low compared to all-NMC makers.

(Also, I have not seen any mention of CATL actually selling NMC811 cells in volume. As far as I can tell, they “have” them the same way the Korean makers “have” them…)

You’re probably right about buses. Most of CATL’s passenger car capacity was added very recently so I have to believe that’s primarily NMC. I agree they probably don’t sell 811 in volume yet.

I know this is pedantic, but I can’t help myself: “catched” is not a word, the past tense of “catch” that you’re liking for is “caught”. Also it’s “fewer GWh”, not “less GWh”, because GWh are countable. Hence “less water” (can’t count water) but “fewer bananas” (because bananas can be counted). My mother is an English teacher; that’s why I catched that 😉

Can’t upvote this enough!

I came to the comments to say just this – it literally made me shudder when I saw “catched”.

Like, more cow bell, not give me 10 decibels more. It can be counted, but it’s inconvenient. You don’t say I need 10 more lumens of light, but just give me more light, because it’s not normally counted, not that a way to count it doesn’t exist. Water can be counted in gallons, or quarts.
Nice catch though.

We don’t count 1 waters, 2 waters, … we count 1 gallon of water, 2 gallons of water. Fewer gallons, less water. (Actually, we do poetically refer to “Land of many waters.”)

Water is made of atoms. Atoms may be counted. It’s fewer water.

If I say I have one water, clearly I’m referring to a single molecule of H2O.

Nope! Atoms may be discrete but they are not countable. Not by you anyway.

On a different note:
Appearing quite frequently here is the use of “amount” as in “amount of people” or “amount of cars produced”. Makes my teeth grate! (or “great” if I want to go down the TrumpleHole without putting the “breaks” on).

Are people so innumerate these days that they have forgotten the word “number”?

and you would be WRONG about not being able to count atoms.
By dropping temps and using lasers to manipulate, along with SEM, it is possible to count small quantities (do not ask for 999,999 atoms of HE).

Agreed, but “catched” was definitely used back in the 1800’s in poetry and other writings. I think more accurately it is now considered an obsolete word no longer used than “just not a word”

Why the English language is somewhat non sensical.

Take catch then caught

Then why is snatch snatched and not

Just sayin’. lol

ain’t was not a word and now is included in the dictionary. Weird.

If cobalt is a concern, Tesla could consider buying such company for china production, they are said to have a production ready line that could make batteries 500wh/kg.

Cobalt is not a concern. Currently purchased from Canada, soon to be eliminated in future versions of the battery.

From who? Where? In “Cobalt, Ontario?”

Panasonic buys NCA powders from Sumitomo Metal Mining, which gets the nickel and cobalt form their mines in the Philippines.

no cobalt is current mined in Canada.

This is basically an advertisement — and it’s full of obvious BS.

First of all, the “ready for mass production” claim seems ridiculous, considering they are also claiming to have achieved mass production of solid state batteries six years ago… Yet nobody uses them? What they actually mean is “ready to solicit a lot of money from investors, in the hope that we can actually put it into production at some point in the future…”

Indeed if their claims were verifiable, Tesla (and surely other car makers as well) would already be doing business with them.

I also call BS on their claim that stacking multiple cells in series in a single housing somehow simplifies the BMS. While it certainly improves packaging (and is actually plausible with fully solid state cells), that doesn’t remove the need for cell balancing.

Last but not least, the density numbers of other architectures they cite for comparison are obviously wrong. Actual research shows that prismatic can cells are somewhat behind pouch cells at pack level, while cylindrical cells are in fact *ahead* of both. (I suspect they counted the entire Model 3 pack assembly, which includes a bunch of components that aren’t actually part of the battery…)

Stacking cells, in “Parallel”, “Might” help simplify BMS, it seems, but in “Serial?” Doubt it.

Putting cells in parallel does not affect the BMS at all. It doesn’t even see them as separate entities.

Cobalt content could be even less in 2018 and that’s why the price of battery / KWh has gone down from $209 to $176.
Will it hit the sweet spot of $100 / KWh in 2021.

Nah, that would only explain a small part of that price reduction. The complete cobalt contents of an old-fashioned NMC111 cell amounts to maybe $40/kWh; going to NMC622 would only reduce that by some $15/kWh.

(Another part of the reduction comes from higher density of the more nickel-rich cathodes — but that still only explains part of the price drops. The rest is from manufacturing improvements, along with general economies of scale.)

> taking into consideration that Panasonic is way above the competitors (see results in 2017), we can draw some assumptions. I love how you’ve clearly meant to write “draw some conclusions”, realised that you can’t actually draw conclusions using 2018 cobalt numbers but 2017 capacity numbers, changed that single word to “assumptions”, and then proceeded to draw conclusions although you’ve kind of admitted that can’t be done. You expect Panasonic to still be #1 in 2018, and implicitly assume the ratio of Panasonic’s output to LG’s will be the same as in 2017. That may be a reasonable assumption to make. I certainly share the expectation that Panasonic will turn out to have made less cobalt-intensive cells than LG in 2018, because they are using NCA chemistry (which allows reducing C). If we assume they use less cobalt per kWh and know they used more cobalt in total, it follows that their output (kWh) was also greater. Why not simply wait until the 2018 numbers are in? If waiting for the numbers isn’t an option, why not simply explain the assumptions and briefly state why you’re making them? On the other hand, admitting that your conclusions are really your assumptions,… Read more »

Maybe the author “Assumed, You Too Would Conclude” the same as they did? Or maybe they “Coluded to Assume?” (And spelled that wrong too? … Maybe I “Catched it All Wrong?”)

1. Panasonic cannot be “lowest” if BYD uses ZERO.

2. Your 2017 chart surely does not include BYD e-bus battery production, only its production for passenger cars. Can you double-check before drawing the 2018 chart?


Does anyone know if any of the cbattery companies is doing anything about sourcing cobalt responsibly? The information “debunking” the videos used to discredit EVs me tions that these videos are from mines that are not only for cobalt or not only for cobalt for EVs, but that don’t really explain what companies are doing to avoid being part of it. I am really interested in a Nissan eNV200 (6-12 month waiting list in Switzerland) but my wife brought this argument up and I didn’t have a coherent rebutal that would really satisfy myself. I looked into AESC’s website but there is no information whatsoever about sourcing their cobal responsibly.