Tesla seeking 56% cut in $/kWh, $25K EV in three years through manufacturing and design innovations
23 September 2020
Elon Musk used Tesla’s much-anticipated Battery Day presentation not to disclose some super cell, but to outline an aggressive, comprehensive and radically innovative approach to battery design, manufacturing and vehicle integration that he projected will result in a 56% reduction in cost/kWh and a $25,000 electric vehicle in about three years.
The baseline precept for the presentation was the centrality and associated difficulty of manufacturing new technologies at a scale sufficient to make a tangible difference. Rapidly scaling global battery manufacturing capacity, Musk pointed out, is absolutely central to achieving large-scale electrification and climate goals.
Tesla has two battery-related goals, Musk said. The first is terawatt-scale battery production. “Tera is the new Giga.” Gigafactory Nevada currently has a capacity of about 0.15 TWh. The second goal is a more affordable battery cell.
What troubles me is that we don’t yet have a truly affordable car. The curve of the cost per kWh of batteries is not improving fast enough.
—Elon Musk
Drew Baglino, Tesla’s Senior Vice President, Powertrain and Energy Engineering, said that the company has a plan to halve the cost per kWh based not on a single innovation, but across the entire chain from raw materials to vehicle integration. 55Musk and Baglino described five major battery manufacturing sub-areas in which Tesla is developing new technology:
- Cell design
- Cell factory
- Anode materials
- Cathode materials
- Cell-vehicle integration.
Tesla projects that its innovations in battery design, manufacturing and vehicle integration will result in a 56% reduction in cost/kWh as well as an increase in range.
Cell design. Musk and Baglino described a major design change for cylindrical cells: a move to a tabless battery with a “shingled spiral” replacing the traditional anode/separator/cathode jellyroll. The new tabless battery enables dozens of connection into the active materials and reduces the electrical path length from 250mm to 50mm.
With the new technology, comes an accompanying move to a larger format: 4680 (46 mm diameter, 80 mm long), compared to the current 2170 and 1865. The new cell gains 5x energy, +16% range, and 6x power just from form factor alone.
Baglino said that Tesla is starting to ramp up manufacturing of the new tabless cell at the company’s pilot 10 Gwh production facility in Fremont; it will take about a year to reach 10 GWh annualized rate.
Tesla attributes a 14% reduction in cost per kWh just to the cell form factor change.
Cell factory. Battery manufacturing comprises four main processes: creation of the electrodes (coating); winding; assembly; and formation (charging for the first time, verifying quality). Tesla is addressing every step in the process with an eye to making it better and more scalable.
The current wet coating process consists of mixing powders with solvents; coating and drying in an oven; solvent recovery; and compression to the final density. Tesla is moving to a dry coating process to eliminate the solvent step, supported in part by the acquisition of Maxwell Technologies, the ultracapacitor company.
[Dry coating] is very hard to do. The dry coating [Maxwell] had was sort of like a cold proof of concept. We have revved the machine four times. It’s not in the bag. There is still a lot of work to do. There are major issues in scaling this up, but it is close to working. It does work, but with not a high yield. There is a clear path to success but a ton of work to get there. We will probably be on machine rev 6 or 7 by the time we get to scaled production.
—Elon Musk
Tesla is also working to redesign the assembly process itself using the bottling industry as inspiration, and focusing on speed and density of the production process itself (i.e., smaller factory footprint).
The aim of Tesla is to be the best in manufacturing of any company on earth.
—Elon Musk
Tesla is taking a similar approach to formation, moving to charging thousands of cells at once. This will result in an 86% reduction in formation investment (which currently accounts for about 25% of the investment in assembly) and a 75% reduction in footprint.
Viewed another way, from a smaller footprint than Giga Nevada, Tesla could get many times the cell output.
Tesla plans to manufacture 100 GWh of cells in 2022, and 3 TWh by 2030. Tesla will continue to use its cell supplier partners. Tesla production will be supplemental to what it buys from the suppliers, reducing the weighted cost per cell. Long term, Tesla sees an incremental 18% reduction in the cost/kWh due to its projected assembly innovations.
Anode. Tesla is proposing a step-change in the capability and cost of silicon anode materials by moving to raw metallurgical silicon, thereby cutting out the expense of the other current highly engineered materials.
Tesla says that it will stabilize the surface of the silicon with elastic on a conductive polymer coating. The silicon material will deliver an anode cost reduction along with an increase in range.
Cathode. Tesla will take a three-tiered approach to the cathode materials in its batteries, Musk said: iron for cycle life, nickel-manganese for an intermediate solution, and a new high-nickel cathode material targeting very long range, the Cybertruck, and so on.
Beyond the metals themselves, the cathode process is a big target for Tesla. Tesla believes it can reduce investment by 66% and the process cost by 76%—and yield zero waste water—by directly consuming raw nickel metal powder, and simplifying the entire process. In other words, Tesla is proposing to eliminate intermediate nickel production.
Tesla will build a new cathode facility in North America, and co-locate lithium conversion at that site, using a new acid-free saline extraction process. Tesla said it has secured a terawatt-hour scale lithium resource in the US. Tesla will also use material from recycled batteries.
Tesla said its cathode innovations will contribute to a 12% reduction in the cost/kWh.
Cell-vehicle integration. Tesla is moving toward a process in which a single-cast front-end and single-cast back end will be joined to a structural battery pack fabricated in a direct-to-pack process—i.e., no intermediate assembly of cells into modules which are then integrated into a pack.
Tesla is already using single-piece casting from front and rear body components. To achieve that, it commissioned the largest casting machine ever made, and also developed a new high-strength aluminum alloy that is very castable.
The structural battery serves a dual purpose as energy device and as intergal structure, very similar to current aircraft design in which the fuel tanks are intergal to the wings.
Structural batteries improve mass and range, Musk said. Also, the construction technique allows cells to be packed more densely. Structural adhesives glue the cells to the top and bottom sheet; the steel shell case of the battery cells transfer sheer from the upper face sheet to the lower. Load is transferred into the pack in a smooth way so that there is no arbitrary point load. The result is even stiffer than a regular car, Musk said.
Longterm, any car not with this architecture will not be competitive.
—Elon Musk
Such a design also enables massive simplification in the factory, with a 35% reduction in floorspace. Tesla calculates this is all worth another 7% reduction in $/kWh, bringing the total reduction to 56%.
It will take a year to 18 months to realize the advantages. Full realization is in three years or about. If we could do this instantly, we would. It bodes well for the future, with long-term scaling massively increased. Long term, we want to make about 20 million vehicles per year.
What does this mean for future products? EV powertrains that cost less than combustion engines. About three years from now, we can make a compelling $25,000 electric vehicle.
—Elon Musk
The volume of the cell is increased by 5.48 and they are getting 5 times the energy and 6 times the power.
IOW it is the same old chemistry in a rather larger cell, with slightly different trade offs.
And after years of proclaiming that small cylindricals were the way to go, they are moving half way to the ease of assembly of the still larger prismatics, however incurring more issues in cooling which are easier to handle in a prismatic.
Musk is fabulous at selling old wine in new bottles.
Posted by: Davemart | 23 September 2020 at 03:52 AM
@Dave, it may be that there is no new wine on the horizon and repackaging it in larger bottles is worthwhile. It is 14% here and 20% there and 4% here etc. He is slowly grinding through the process of making cheaper EVs and cheaper batteries, and getting to grips with the scale that things will have to be done at.
Imagine what will happen when we have $25K cars (lets say in 4 years time). Very rapidly, the wealthier countries will transition to EVs and by (say) 2035, 1/2 - 2/3 of the cars on the roads will be EVs.
You'll have a lot less pollution, but just as much congestion, and the price of oil will be on the floor.
Posted by: mahonj | 23 September 2020 at 04:43 AM
Tesla used 18650 cells because they were the cheapest commodity cells at the time.
Those with stardust in their eyes proclaimed it was because it was the bestest best best technology, and they had a massive lead over other car companies because of it.
I was even told by engineers who were star struck that it did not make assembly more complicated and expensive.
Well, Tesla don't agree. and have now moved towards saying that fewer makes life easier, which was blindingly obvious in the first place.
Posted by: Davemart | 23 September 2020 at 07:54 AM
Davemart,
I don't think the new 4680 format shown in the conference is "prismatic". They look pretty cylindrical to me.
Tesla cylindrical, cooled cells have the best lifespan track record on the industry. All your talk around it is nonsense. BYD prismatics are comparable, but much heavier.
The energy density of the new cell has not been clearly explained in the conference; maybe there is a tradeoff removing Cobalt. Maybe the energy density is really better, but at pack level that improvemente is reduced (due to the heavy cooling systems Tesla uses).
Posted by: peskanov | 23 September 2020 at 08:01 AM
There was nothing in the Battery Day presentation we did not already know, just the impact of how the several changes impact battery cost. These are critical developments particularly to a manufacturing engineer like myself.
We have a general understanding of what the energy density of the new 4680 cells will be based on current technology and the impact of using silicon anodes, thick cathodes, and dry processing (probably a 25-40% improvement on current cathode metals (LFP, LMNO, and LMNC). So no 400-500 Wh/kg battery yet.
The most interesting thing is recognition that LFP will be part of the Tesla plan (already used in the Made in China Tesla 3 ).
The impact of LFP will happen soon, just look at current off the shelf batteries like the Guoxuan cylindrical 33x131 batteries which have 178 Wh/kg (https://www.shanghai-electric.com/group_en/upload/resources/file/2019/11/13//80901.pdf). Add the Tesla improvements and a $25k EV is definitely possible.
Posted by: Account Deleted | 23 September 2020 at 10:01 AM
@peskanov
I never said that they are prismatic, I said that prismatics are bigger again, as well of course as a different shape.
And the no improvement in energy density is at the cell level, not 'really better' but reduced at the pack level.
The figures given were for the cells.
Posted by: Davemart | 23 September 2020 at 12:10 PM
Davemart,
the secret sauce which provided excellent lifespan to Tesla batteries is perfectly known at this point: keeping the cells cool. Not just ambient temperature, but cool, south of 30 degrees Celsius which ages any Lithium cell fast.
And this needs small cells to avoid heat pockets. Lithium-ion materials are not good heat conductors.
This was told to me by a Tesla engineer himself, years ago, and I think is common knowledge by now.
The other factor they were trying to avoid was electrode paste movement inside the cell. For example, in vertical pouch cells the paste tends to fall slowly to the bottom on each cycle. Cylindrical cells behave much better in that respect.
So yes, of course they know bigger cells are cheaper to assemble, but robustness was the priority at it's obvious to me their bet payed off. Other early electric cars like the Leaf killed their battery in less than 8 years, I am bored of hearing bad reports about them. I wish they would have used the cylindrical format.
The 2170 was Tesla's compromise between "can be cooled efectively" and "as big as possible".
The reason of the bigger new format (4680) was explained yesterday: the tabless cell format they have created has less electric resistance, a lot less, therefore it produces less heat on charge or discharge. That allows a bigger cell to be kept cool as previous packs.
Posted by: peskanov | 23 September 2020 at 02:31 PM
There is a major improvement in EVs that has gone unnoticed by most viewers of the Battery Day event. Elon said that the tabless cell architecture "is way more important than it sounds" and he is right. What is the general public's biggest gripe about EVs? It is charging time versus gasoline fill-up time. During Battery Day, Elon briefly showed a chart demonstrating that, due to the lower internal resistance of the tabless architecture, the 4680 cell will charge as fast as the 2170 cell. Given that the cell holds 5x energy over the 2170, supercharging your car could take one fifth the time it does today, once the superchargers are upgraded to handle the increased rate of power delivery. That is truly game changing and eradicates one of the biggest complaints of ICE car owners- charging time on a trip. Adding 300 miles of range could take only 15 minutes- 100 miles could happen in 5 minutes or less. Complaints about range and battery life are old news. Now, the advantage of gas stations will be gone too!
Posted by: Michael45433397 | 23 September 2020 at 03:45 PM
@peskanov
The 'secret sauce' in Tesla's batteries was the one which made them class leaders in battery fires.
Of course that was not just because they had somehow jimmied together toothbrush batteries which they used because they were cheaper and then worried about cooling later, but they also were as obsessed as a 12 year old about acceleration figures etc at the expense of safety, to which they simply paid lip service.
To cool a battery it is far easier to use prismatics which can use cooling plated top and bottom than the Heath Robinson cooling tubes Tesla uses..
They are making cooling problems worse by moving to the new bigger battery cells, which cylindricals really don't like and may be the reason behind the slight energy density DECREASE , but no doubt they will fudge something up , then declare it to be groundbreaking genius.
Posted by: Davemart | 24 September 2020 at 02:14 AM
Davemart,
"Tesla leader in battery fires"?
Lol, now it's obvious you are just trolling. There has been much more fires in pouch and prismatics than in cylindrical batteries. The number of fires in prismatics is less known, as the technology is mostly used in china, but still many reports exist on cars and buses burning.
Prismatics are usually pretty thick, around 50 mmm. For thinner cells they usually choose pouch formats, and there are very few examples of them being cooled. The Chevy Bolt is one of them, and the (few) existing battery degradation graphs show faster aging than Tesla packs.
I see your idea of Tesla engineering is that they don't know anything about batteries; I find this pretty fun.
Posted by: peskanov | 24 September 2020 at 06:00 AM
@peskonov
If you think what you are saying is true, then you are entirely ignorant on the subject.
If you don't, then it is you who are trolling.
Here is a list of Tesla fires:
https://www.thedrive.com/news/28420/parked-teslas-keep-catching-on-fire-randomly-and-theres-no-recall-in-sight
I look forward to your lists of the fires you claim in pouch and prismatic battery cars.
That might be difficult, as they are rare to non existent.
Posted by: Davemart | 24 September 2020 at 09:03 AM
I will concede that there are fewer battery fires in other car brands, but I don't see that as significant. Yes, there is only 4 or 5 Hyundai Kona fires, 1 Taycan fire, 1 Panamera fire, 1 eGolf fire, 1 or 2 Fisker Karma fires,and the list goes on in that fashion...
However, how is that significant? All these cars have sold a fraction of what Tesla sells. 8 Tesla fires versus all the cars sold in this list hardly seems worse.
In the world of cheap electric motorcycles, fires are pretty common, and nearly all of the are pouch or prismatic.
I can accept Nissan having a better track record in security (which imo is largely offset by the horrible tales of battery degradation; Nissan is hardly a good example of good EV engineering), but I am pretty sure this not due to cell format, but better electronics.
In my experience (and yes, unfortunately I have first hand experience) a parked EV burns when overcharged. Any cell going north of 4.3V starts heating.
Now, when you overcharge a pouch or a prismatic, it heats, swells, punctures and starts emitting gas at fast speed. If the cell is enclosed, the gas ignites very easily.
I have seen this several times performing tests.
A very similar process happens with prismatics, although they swell less dramatically and usually puffs less.
Guess what? The only format which offers a [b]passive protection[/b] against overcharging is the cylindrical format. An overcharged cylindrical cell contains pressure to a limit, and then breaks electric contact using a pressure valve. This is called a CID.
All of the cylindrical cells I have tested for overcharging did correctly break the circuit. The only way I have caused cylindrical cells to fire is puncturing them using a nail, and even then, only those rated for high discharge would catch fire.
Posted by: peskanov | 24 September 2020 at 11:45 AM
It's my hope that Tesla will sell some of their cells to the aftermarket so small equipment makers have better stuff to use for projects, i.e., riding lawn mowers, car and truck conversions, DIY projects, etc. Perhaps someone will even rescue some of the 500,000, plus perfectly good used Nissan Leafs that will be in want of an affordable replacement battery.
Posted by: Lad | 24 September 2020 at 07:53 PM
@peskonov
So you concede that your original claim that:
'There has been much more fires in pouch and prismatics than in cylindrical batteries.'
is simply untrue, which is false.
You then go on to seek to dismiss the data, on the grounds that there are more Tesla cars.
They only moved into such a heavy preponderance with the advent of the Model 3, which as they are mostly almost brand new one would hope that they would have very few battery fires, which are usually exacerbated by age etc.
There are plenty of Nissan Leaf and BMW i3s etc to compare with the Model S and X, with only the latter two having a pronounced proclivity to become a barbecue.
The issue may or may not have been dealt with in the Model 3, it is too early to tell, but far from my 'trolling' as you claimed, my statement about fire issues in Tesla cars are no more than the truth.
They are far more interested in 0-60 times than safety, whatever lip service they pay to it.
Posted by: Davemart | 25 September 2020 at 08:46 AM