The Promise of Maxwell

[Cosmacelf]

The 16X production capacity density increase is a big deal for the GigaFactory. That would free up a lot of factory space for other production and increase cell output.

BTW the original 85KWH cells had an energy density of 240WH/Kg and tear downs of Model 3s found the 2170s in use today are 247Wh/Kg, so the 300 number probably does include the Maxwell tech. The dry electrode batteries are probably a bit less dense, so on the pack level we won't see a 50 WH/Kg increase, we'd need to know the weight of a dry electrode 2170 to get an estimate of that.

A battery cell that has 1000 WH/Kg, but the density of styrofoam would make a pretty low energy density EV pack.

Why do you say the dry electrode battery cells are probably less dense?

 

[Doggydogworld]

The energy density numbers are also telling given Musks recent tweet (June 30th):

FWIW, based on calcs I did 10 years ago, cross-over point for Li-ion beating kerosene is ~400 Wh/kg. High cycle batteries are just over 300 Wh/kg today, but probably exceed 400 in ~5 years.
​

According to Maxwell they had demonstrated 300 Wh/kg in January and Musk backs this up. (Unless he is talking about Tesla battery tech prior to Maxwell, I am not informed enough on batteries to know this)

Musk could be talking about several things. CATL claims 300 Wh/kg at the cell level. Maxwell talks about thick electrodes which increase specific energy at the cost of reduced specific power. 300 Wh/kg with a 1C rate limit wouldn't work for EVs, of course (made-up example to illustrate the point).

Ok, I’ll bite. What context have you gained from doing this?

Maxwell has talked about dry battery electrode coating for over 10 years. They did deals with JCI/Saft and Lishen ages ago, but nothing happened. They've hawked it to pretty much everyone else with no takers. The 500 Wh/kg claim amounts to "bring us a 500 Wh/kg chemistry and we'll adapt our process to coat electrodes with it", but they let people think they've got some magic formula.

There's a slide somewhere where Maxwell, in a fit of modesty, claimed their process could save $200-300 per BEV battery pack. With Tesla's scale that easily justifies $200m of stock, but it's not the revolutionary leap forward people are fantasizing. It brings to mind all the claims that 2170s would improve energy density 35%. Actual improvement..... ~0%.

 

[Cosmacelf]

Maxwell has talked about dry battery electrode coating for over 10 years. They did deals with JCI/Saft and Lishen ages ago, but nothing happened. They've hawked it to pretty much everyone else with no takers. The 500 Wh/kg claim amounts to "bring us a 500 Wh/kg chemistry and we'll adapt our process to coat electrodes with it", but they let people think they've got some magic formula.

There's a slide somewhere where Maxwell, in a fit of modesty, claimed their process could save $200-300 per BEV battery pack. With Tesla's scale that easily justifies $200m of stock, but it's not the revolutionary leap forward people are fantasizing. It brings to mind all the claims that 2170s would improve energy density 35%. Actual improvement..... ~0%.

OK, thanks for the info.

BTW, to be fair to Tesla, Tesla never claimed the 2170s would be revolutionary. IMHO they moved to the 2170s because it was cheaper than staying with the 18650s - it was just a cost savings move, which is fundamentally what all this battery cell technology is aiming to do: reduce cell cost relative to gravimetric energy density (and yes, I'm including longevity in there too. If you have a longer lived cell, all that means is you can work it harder for the same battery life as your old cell, so that translates into cheaper cells).

 

[Doggydogworld]

OK, thanks for the info.

BTW, to be fair to Tesla, Tesla never claimed the 2170s would be revolutionary. IMHO they moved to the 2170s because it was cheaper than staying with the 18650s - it was just a cost savings move, which is fundamentally what all this battery cell technology is aiming to do: reduce cell cost relative to gravimetric energy density (and yes, I'm including longevity in there too. If you have a longer lived cell, all that means is you can work it harder for the same battery life as your old cell, so that translates into cheaper cells).

Right, I didn't mean to imply Tesla claimed 35% energy density improvement for 2170s. It was bulls taking a few generic comments about energy density improvements since the Roadster and turning off their brains. Same with Maxwell. Tesla has said very little but some are taking Maxwell's PR and making goofy extrapolations.

 

[wdolson]

Why do you say the dry electrode battery cells are probably less dense?

Liquid electrodes have a density close to that of water. Most of the dry materials that could be used for electrodes have a lower density. This would contribute to a lighter cell overall. However, probably not significantly, maybe 10%?

 

[brucet999]

Liquid electrodes have a density close to that of water. Most of the dry materials that could be used for electrodes have a lower density. This would contribute to a lighter cell overall. However, probably not significantly, maybe 10%?

Am I confused here? I understood from reading about Maxwell's dry electrode process that it saved having to evaporate liquid solvents out of a slurry, then collect and re-use them. I got the impression that the result of driving off solvents left dry electrode materials behind in the finished battery cell. If that is the case, there would be no weight difference in the two types of cells, assuming the electrode material is the same chemistry.

 

[Hugh-SG]

Cites, please. I know plenty of oil folks. None of them are fans of selling gasoline - they'd all much rather sell NG - which is how most most EVs are powered...

Sorry but 96% of of the electricity where I live comes from hydro electric dams. Thus EV's in my area are Green Power fueled. NG is still a pollution product, but it beats the heck out of wood or coal burning.

In 2016 whilst visiting Arizona, I saw billboards and TV ad's pleading for voters to save 23,000 Arizonian jobs by Not allowing the sale of Arizona"s solar power farms to "Big Oil".

Could it be those smart folks from from the desert state remember how electric trolley buses and electric suburban trains were killed off by "Big Oil" and the tire companies and they do not want to see a repeat.

Food for thought... Cheers.

 

[neroden]

There's a slide somewhere where Maxwell, in a fit of modesty, claimed their process could save $200-300 per BEV battery pack. With Tesla's scale that easily justifies $200m of stock, but it's not the revolutionary leap forward people are fantasizing. It brings to mind all the claims that 2170s would improve energy density 35%. Actual improvement..... ~0%.

$300/battery pack would be about $300 million at the scales Tesla is planning in the near future, and *not* enough to justify $200 million in stock, since the stock price will go up and it'll effectively be a larger cost.

For Tesla, however, the savings in being able to use *less factory space per line* and *much quicker line setup* probably end up being worth far more money than that. It speeds up the process of ramping up battery production if you don't have to build as many gigantic ovens and solvent recovery systems and you don't have to go through the whole process of introducing the toxic solvent in the first place...

I think we should basically assume that these batteries are as good as the wet-electrode battteries, with slightly lower operations costs (no solvents), somewhat quicker production (no drying time), much lower capital costs (much simpler equipment), and much quicker installation.

Since Tesla is repeatedly constrained by how fast they can ramp up battery production, this makes the tech worth multiple billions of dollars. If they can spin up a line six months or a year faster because they don't need to build an extra block of building, it improves their entire ramp up. If they can reduce the time from raw materials entering the factory to packs using those materials coming out, it's essentially like shortening the delivery time for cars -- it improves working capital.

So yeah, the resulting batteries can be about the same as the existing ones, and the marginal cost of production can be only slightly lower, but the effect on Tesla's financials can be enormous by improving working capital and reducing the amount of capex needed to scale up.

 

[wdolson]

Am I confused here? I understood from reading about Maxwell's dry electrode process that it saved having to evaporate liquid solvents out of a slurry, then collect and re-use them. I got the impression that the result of driving off solvents left dry electrode materials behind in the finished battery cell. If that is the case, there would be no weight difference in the two types of cells, assuming the electrode material is the same chemistry.

The way li-ion cells are made today the electrolyte is wet even in the final product. That's why they can catch fire, if the cell gets crushed, the electrolyte can get squeezed out, then the electrodes touch and energy discharges very fast generating a runaway heating reaction. With a solid state electrolyte the electrolyte can't get pushed out of the way and the chances of the electrodes touching is much smaller.

There are other materials in li-ion cells that are deposited on a surface and must be cured for 24 hours or more.

 

[Doggydogworld]

The way li-ion cells are made today the electrolyte is wet even in the final product. That's why they can catch fire, if the cell gets crushed, the electrolyte can get squeezed out, then the electrodes touch and energy discharges very fast generating a runaway heating reaction. With a solid state electrolyte the electrolyte can't get pushed out of the way and the chances of the electrodes touching is much smaller.

There are other materials in li-ion cells that are deposited on a surface and must be cured for 24 hours or more.

This is mostly right, but "electrodes touching" can happen with any type of electrolyte in crush/puncture/etc. events. The problem is the liquid electrolyte typically used in Li-ion cells, hexafluorophosphate (LiPF6), is extremely flammable. Things go downhill once a fire starts, because the cathode feeds oxygen to the fire. That's why fire extinguishers don't work -- it keeps burning even if you blanket the cell and cut off all outside oxygen. Tesla uses Li(NiCoAl)O2 (NCA) cathodes with a pretty weak O2 bond that releases oxygen at relatively low temperature, further reducing thermal stability.

Solid electrolytes tend to be much less flammable and mostly avoid these problems. But many solid state cells use metallic lithium anodes, which reacts with water (including water vapor in the air) to produce H2 gas which is certainly flammable. There is always some risk when you contain a lot of energy in a small space.

 

[gluu]

There is likely an energy density improvement to 300Wh/kg. For example, Elon mentions maybe they can get the semi to 600mi range, up from 500. This is the same percentage improvement going from 250Wh/kg -> 300Wh/kg would create. Elon also said at autonomy day about having million mile batteries next year, which is a roughly 2 - 2.5x increase in battery life. It's too coincidental IMO for it to not be the Maxwell tech. At least from the published papers by Maxwell, these are the claims being made and they all dovetail in nicely with the hints that Elon gives. The dry electrode batteries can also take higher C rates without degrading as well according to Maxwell, so we should also see improvement in charging times.

Keep in mind, if you read more of Maxwell's quarterly reports, Tesla started working with them sometime in 2016, quite a long time ago. Over the years in earnings calls you can see that they reached milestone after milestone, first with "proof-of-concept" milestones, and in 2018 they even had a pilot line for electrode manufacturing which looked successful. There is only discussion of a single OEM working with them trying to apply their tech, and I think it's quite obvious after the fact that it was Tesla. To me, it looks like Tesla has already de-risked the Maxwell acquisition being BS. The last step looks to be going to mass production.

I believe in 2018 Q1/Q2 Maxwell Earnings, Maxwell discussed the "leading automotive OEM" they were working with on brought in their own chemistry to try with the dry electrode technology, Tesla actually brought in their cobalt free chemistry, which Elon has also hinted at.

I think in 2020 we will probably see Cobalt Free, 300Wh/kg, reduced capex resulting in 10-20% reduced costs, and also cutting Panasonic's guaranteed 10% net income, here's my estimate for cell costs, assuming currently $100/kWh

$100/ kWh, density increase to 300Wh/kg assumes for same cost, can get 20% more kWh per dollar, so 100/1.2 = $83.3/kWh
$83.3/kWh, decrease in capex for ovens/solvent/recovery, resulting in 10% cost savings, so 83.3*.90 ~= $75/kWh
$75/kWh, cutting panasonic's guaranteed net margin of 10%, so 75 - 7.5 = 67.5.
some cost reductions for removing cobalt, and the net margin of lithium suppliers (Elon mentioned going into mining, this is almost certainly for lithium), assume 0 for "margin of safety" i.e. reality is more difficult than theory.

Costs are not likely to be this low initially, but theoretically we're talking about a >30% decrease in cell costs for Tesla going forward with the ability to scale to really large numbers in terms of GWh.

At some point with this tech it probably just makes sense to drop the price of the Model 3 and make a ridiculous number of vehicles.

 

[neroden]

Nitpicking:

Betcha the cobalt-free chemistry isn't actually cobalt-free; it's practically impossible to get pure nickel without some cobalt contamination. But the problem was always *how much* cobalt was needed; if they're down to the point where they're trying to remove excess cobalt impurities, it's close enough to cobalt-free.

 

[ralph142]

Nitpicking:

Betcha the cobalt-free chemistry isn't actually cobalt-free; it's practically impossible to get pure nickel without some cobalt contamination. But the problem was always *how much* cobalt was needed; if they're down to the point where they're trying to remove excess cobalt impurities, it's close enough to cobalt-free.

isnt it the need to add cobalt, and the nastiness of the sourcing thats the issue, not cobalt inherent to the nickel ?

 

[wdolson]

isnt it the need to add cobalt, and the nastiness of the sourcing thats the issue, not cobalt inherent to the nickel ?

The use of cobalt is more political than an engineering problem. The largest source of cobalt in the world is the Congo and they essentially use slave labor to mine the stuff. Though I never thought it was a serious problem, there is plenty of other cobalt in the world. I think the 3rd largest deposit is in Cuba in developed mines that have been shut in for 60 years. Australia and Canada have some cobalt too.

 

[ralph142]

The use of cobalt is more political than an engineering problem. The largest source of cobalt in the world is the Congo and they essentially use slave labor to mine the stuff. Though I never thought it was a serious problem, there is plenty of other cobalt in the world. I think the 3rd largest deposit is in Cuba in developed mines that have been shut in for 60 years. Australia and Canada have some cobalt too.

There an old closed mine in California (?) iirc as well. But again, we’re talking about the need to add cobalt, not the nickel purity, as I understand it.

 

[DrJohnM]

I probably get down voted for this, but one thing people need to consider is the price Tesla paid for Maxwell. Say about 200-250M. That's really not a lot. Suggests to me that Maxwell tech is not some holy grail.

But it is obviously important to some extent to Tesla's strategy. And what exactly that is remains to be seen.

I think everybody (like Galileo Russell) going on about how Maxwell is going to transform Tesla is maybe reaching a bit. Personally I'm taking a wait and see attitude. If Maxwell really was some revolutionary breakthrough tech, there would have been a major bidding war and it's price would have been over 1 billion, considering how important battery tech is to a future trillion dollar auto industry. Either that, or everybody else, including Maxwell, is not seeing what only Tesla is seeing -- a distinct (but unlikely) possibility. Let's see what happens...

If you are only building 10k EV’s a year, then $250k is $5k per car over 5 years. Say you can take $5k off the pack price due to better energy density etc, you only break even. The cost of the tech is only compelling if you are making lots of EV’s and you have to believe it will get you those savings.

 

[Stars aligned]

So what savings brings Maxwell technology to Tesla?
Let's assume Maxwell cells will be produced with 300 Wh/kg. Thats 120% of current 250Wh/kg. If current costs are 100 USD / Wh and production costs stay the same per kg, new production costs are 83 USD / Wh due to increased energy density.
Now production will get even cheaper due to 16 fold less space on the electrode side and other efficiency gains. Let's assume that's another 16%. New production costs at cell level will then be 70 USD / Wh.
35 Gwh production per year with old technology would cost 3.5 bn. With new technology 2.45bn, savings 1.05 bn.
54 Gwh 5.4 bn vs. 3.78 bn, savings 1.62 bn.
Or cells for a 70 kWh Model 3 cost now 7000 USD, then 4900 USD.

It's a pretty good investment, it seems.

Edit: and not yet factored in longer lifetime of the battery (million mile battery, anyone?)

 

[jbih]

Well, I just remember J B Straubel and Elon Musk saying that they ignored all PowerPoint presentations about batteries, and they would only buy a battery "breakthrough" if they could see an actual working cell (and presumably the lab that made it). So they clearly did that with Maxwell. Which means they are fairly advanced on the tech and fairly ready to apply it to production.

There is also the secret floor in Gigafactory 1 which could house experimental battery production.

 

[jbih]

side note: I think only 2 automakers are using cells. Tesla and Rivian.
Everyone else using pouches.

Only Tesla bold enough to make the batteries they need.
All other OEMs want Battery Companies to do R&D and all capital expenses for battery factores/production.

Kreisel Electric is an Austrian company making battery packs with cylindrical cells. And Rimac uses cylindrical cells as well.

 

[brucet999]

There is also the secret floor in Gigafactory 1 …

Is that at GF1 annex, located in area 51?