Will Silicon batteries be a major game changer?

[peter]

Hydrogen is 15 years behind in spite of a 15+ year head start. So this will tell you a lot about the S curve in the relative performance of these technologies. By the time hydrogen is viable, the performance/cost of Li batteries will be many times better than today.

There are many paths to improvement in Li batteries. On cathode, there are chemistry improvements, the most promising is adding a little manganese to LFP for LMFP batteries. On anode you will see increased silicon content in anode material. There can also be improvements in binders and additives that will provide incremental performance improvements. One of the most promising areas of improvement is the separator material, where new separators are being designed that inhibit dendrite growth, which allows chemistries to be pushed even further. Changes to cell format can also make a difference in cost, like Tesla 4680 battery, or large format prismatic cells like BYD blade. There are improvements at the pack level, like blade battery pack and other cell to pack technologies. Finally, there is a lot that can be gained from new manufacturing process, like Tesla dry coating process or semi solid battery process.

The Li battery industry is still very young. Li batteries were only invented in the 1980s. Compare this to silicon transistor, which was invented in 1950s. Batteries don't have as fast of an improvement rate as transistors, so we will not see a 1000x performance/cost improvement over the next 30 years, but I would bet a 10x improvement is very possible. In the same time, I doubt hydrogen will even have a 2x improvement.

Boy, even more optimistic than I am about lithium batteries, must not work in the industry!

I jest, I jest, I have those 46xx cells (and even larger cells) and large blade cells to play around with and all that new technology is definitely exciting. But 10x improvement? Not going to happen. Hard physics cap, Faraday's constant/Li atomic mass, 13901 coulombs per gram, theoretical cap of 11.6kWh per kg of lithium. That's not factoring the intercalation compounds of the cathodes and the anode, nor is it including the housing, separator films, etc. Graphite as an anode has a theoretical cap of 372 mAh/g, silicon can potentially triple that, but again, once factoring in all the rest of the build of materials, even doubling current leading edge lithium batteries in energy density is going to be extremely challenging. Even that link Arthur posted, there's a lot of marketing fluff in there about their proprietary silicon anode's 3250 mAh/g potential, but their actual stated performance is about 20% better than current lithium batteries and their stated goal is 40% better over the next decade. Even the 4680 cells, it improved energy density on the cell level from 270wh/kg to about 280-290wh/kg. Incremental improvement. BYD blade and CATL qilin, same thing, small incremental improvements, roughly 180wh/kg because it's using LFP for blade, qilin's around 260wh/kg last time I recall. Averaging about 60wh/kg improvement per decade for the past three decades, which is still phenomenal. The bigger shocker is the cost reduction per kwh, went from over $1000 in 2010 to around $80 right now, but talking to sodium and lithium battery manufacturers, the cost reductions hereon out will be incremental, lithium and the other active material costs don't make up as big of a proportion of the total cost and the outlook on long term raw material availability is a bit murky as well with the explosive expansion in cell production capacity.

It's clear lithium batteries is ahead for certain applications, I don't see a hydrogen car being practical or necessary at all. But I don't want to completely disregard the potential of not just hydrogen, but fuels that can be synthesized with hydrogen, such as methanol, ammonia or synfuel. They're incredibly useful feedstock for many industrial processes and practical fuels for heavy machinery, equipment and transportation vehicles, such as bulk cargo carriers, container ships and potentially aircraft, vehicles that lithium batteries realistically have no chance of penetrating in our lifetimes other than very short haul specialty cases. But for me, besides the carbon reduction or carbon neutral aspect, it drastically reduces SOx, NOx and other particulate emissions, stuff that really matters to port cities, seeing as to how bad the air quality is around LA and the bay area, stuff that actually matters to ones health. Sabatier, Haber, Fischer Tropsch processes are all mature on an industrial scale, these gases and fuels can currently be made cheap using fossil fuels, but if we can deploy massive cheap renewables online, far better to use the energy to produce these fuels than just charging up batteries. Hydrogen infrastructure, I'll be the first to admit is too costly outside of centralized production centers, but methanol infrastructure and ammonia infrastructure is quite doable.

Or in my personal use case, I have a 120kWh battery for my house. It can only store 120kWh, and since I made the pack from cells directly procured from China, it was drastically cheaper than any off the shelf ESS. But at the end of the day, it can only hold 120KWh and takes up quite a bit of space. With my home solar setup, it produced a little over 60kWh a day after the winter solstice a few weeks ago, and can produce about twice that in the summer. But my use in the summer is at most 40kWh a day whereas my use in the winter is 50kWh a day. Adding another 120kWh of storage, even at $80/kwh is not exactly cheap, and will only get me 2 more days of winter. About $300/kg for hydrogen tank, 30kg would yield nearly 1000kWh of electricity and would cover the worst weather that I can possibly fathom. I have room, it'll be stationary so don't need highest compression rate or weight saving materials, could be fitted with cheaper compressor/tank, and if I really wanted to say screw it with efficiency, just pipe the hydrogen into a ICE generator and get cheap backup power. Or maybe I'm just being delusional!

 

[count]

10x improvement in cost performance at pack level, not straight performance, and over a period of 30 years.

Performance has a hard cap although there is currently a lot of room for improvement, but the materials cost can be greatly reduced. The manufacturing processing cost can also be greatly reduced.

I actually do work in battery industry.

 

[saimali]

Boy, even more optimistic than I am about lithium batteries, must not work in the industry!

I jest, I jest, I have those 46xx cells (and even larger cells) and large blade cells to play around with and all that new technology is definitely exciting. But 10x improvement? Not going to happen. Hard physics cap, Faraday's constant/Li atomic mass, 13901 coulombs per gram, theoretical cap of 11.6kWh per kg of lithium. That's not factoring the intercalation compounds of the cathodes and the anode, nor is it including the housing, separator films, etc. Graphite as an anode has a theoretical cap of 372 mAh/g, silicon can potentially triple that, but again, once factoring in all the rest of the build of materials, even doubling current leading edge lithium batteries in energy density is going to be extremely challenging. Even that link Arthur posted, there's a lot of marketing fluff in there about their proprietary silicon anode's 3250 mAh/g potential, but their actual stated performance is about 20% better than current lithium batteries and their stated goal is 40% better over the next decade. Even the 4680 cells, it improved energy density on the cell level from 270wh/kg to about 280-290wh/kg. Incremental improvement. BYD blade and CATL qilin, same thing, small incremental improvements, roughly 180wh/kg because it's using LFP for blade, qilin's around 260wh/kg last time I recall. Averaging about 60wh/kg improvement per decade for the past three decades, which is still phenomenal. The bigger shocker is the cost reduction per kwh, went from over $1000 in 2010 to around $80 right now, but talking to sodium and lithium battery manufacturers, the cost reductions hereon out will be incremental, lithium and the other active material costs don't make up as big of a proportion of the total cost and the outlook on long term raw material availability is a bit murky as well with the explosive expansion in cell production capacity.

It's clear lithium batteries is ahead for certain applications, I don't see a hydrogen car being practical or necessary at all. But I don't want to completely disregard the potential of not just hydrogen, but fuels that can be synthesized with hydrogen, such as methanol, ammonia or synfuel. They're incredibly useful feedstock for many industrial processes and practical fuels for heavy machinery, equipment and transportation vehicles, such as bulk cargo carriers, container ships and potentially aircraft, vehicles that lithium batteries realistically have no chance of penetrating in our lifetimes other than very short haul specialty cases. But for me, besides the carbon reduction or carbon neutral aspect, it drastically reduces SOx, NOx and other particulate emissions, stuff that really matters to port cities, seeing as to how bad the air quality is around LA and the bay area, stuff that actually matters to ones health. Sabatier, Haber, Fischer Tropsch processes are all mature on an industrial scale, these gases and fuels can currently be made cheap using fossil fuels, but if we can deploy massive cheap renewables online, far better to use the energy to produce these fuels than just charging up batteries. Hydrogen infrastructure, I'll be the first to admit is too costly outside of centralized production centers, but methanol infrastructure and ammonia infrastructure is quite doable.

Or in my personal use case, I have a 120kWh battery for my house. It can only store 120kWh, and since I made the pack from cells directly procured from China, it was drastically cheaper than any off the shelf ESS. But at the end of the day, it can only hold 120KWh and takes up quite a bit of space. With my home solar setup, it produced a little over 60kWh a day after the winter solstice a few weeks ago, and can produce about twice that in the summer. But my use in the summer is at most 40kWh a day whereas my use in the winter is 50kWh a day. Adding another 120kWh of storage, even at $80/kwh is not exactly cheap, and will only get me 2 more days of winter. About $300/kg for hydrogen tank, 30kg would yield nearly 1000kWh of electricity and would cover the worst weather that I can possibly fathom. I have room, it'll be stationary so don't need highest compression rate or weight saving materials, could be fitted with cheaper compressor/tank, and if I really wanted to say screw it with efficiency, just pipe the hydrogen into a ICE generator and get cheap backup power. Or maybe I'm just being delusional!

 

[peter]

That's still quite an aggressive target, but then again, 30 years is a long time!

 

[count]

That's still quite an aggressive target, but then again, 30 years is a long time!

There is a tendency to overestimate progress in the short term, but underestimate it in the long term. I have no doubt that a 10x improvement in 30 years is achievable. We are only in the 3rd inning when it comes to the development of Li battery technology.

I am directly involved the the development of new manufacturing processes and technologies that we are validating on pilot lines today. As an example, dry coating and semi solid processes will take 40-50% out of the processing cost and, while there are plenty of challenges, I am quite certain that it will be commercialized within a few years.

 

[Memphremagog]

It's been a year since this thread was started, and a lot has happened. The healthy skepticism in some of the earlier posts (about things like swelling) now has been met with real progress demonstrating that issue has been successfully addressed.

Group14 Technologies saw its JV plant in Korea with SK get fully commissioned, at 2,000 metric tons/yr, 10 GWh. That makes it the first "EV-scale" silicon anode materials plant in the world.
https://group14.technology/resource...livers-scc55-to-over-100-customers-worldwide/

Then, it was revealed that ATL has been supplying silicon anode batteries to Honor, now in millions of Chinese smartphones.
https://www.group14.technology/reso...14s-scc55-powers-honor-magic7-pro-smartphone/

At the end of the year, more battery companies started showing up with very interesting performance.
https://spectrum.ieee.org/silicon-anode-battery-2670396855
https://www.inobat.eu/newsroom/test...on-for-the-high-performance-mobility-markets/

Even end-user case studies are starting to appear:
https://www.electrive.com/2024/12/10/archer-commercializes-air-taxis-in-abu-dhabi/
https://flyingbasket.com/blog/news-...shore-using-flyingbasket-heavy-lift-drones-25

And my personal favorite is the McMurtry Spéirling Pure hyper car. (sorry, they are only making 100 of them, and they cost ~800,000 Euros!).

2025 will see this technology start to go very mainstream.

 

[count]

Yes and in the last year battery manufacturers have dropped prices by about 50%, so that's already a 2x improvement in cost/performance in just 1 year - that is coming just from material cost reductions (thrifting on the amount of copper/aluminum in foils + lower commodity prices), economies of scale, and better quality/cost management. Only 5x to go in the next 29 years. That should be easily possible between improved chemistries, commercialization of new manufacturing process, further material cost reductions, and further pack level improvements.

 

[ak0906]

Enovix (ENVX) is another battery company making progress with silicon-anode cells. Their target end market is consumer electronics - smartphones & AR/ VR. ENVX has signed Joint Development Agreement with 2 of the largest smartphone makers, one US based AR/ VR player has funded tooling for batteries. Key will be High Volume Mfg (HVM), expected 2H 2025. Factory is in Penang, Msia. For HVM ENVX hired guys with semiconductor experience - Raj Telluri & Ajay Marathe (QCOM, MU, AMD experience). CTO is from ATL, Samsung SDI.

 

[Arthur Hanson]

Ultimate game changer? No. Not even remotely close.

Even sodium ion batteries, of which I have quite a few sitting on my desk and are currently in production in commercial volume quantities, it's not remotely close to making a big impact, at least not in the near future.

There are plenty of technologies that offer pathways to lower costs, better performance and longer life. The problem is just as Tanj pointed out above, scaling it up!

Look at CATL, LG Chem, BYD, Samsung SDI, SK On and Panasonic. How much money have they invested in production capacity currently for the respective chemistries they produce? Multi-billions of dollars.

When I visited the battery factories late last year discussing sodium ion batteries, the main reason the Chinese were pursuing it was to reduce reliance on imported materials when making lithium battery variants. The local provincial governments were subsiding costs massively to just bring it to parity with LFP and NMC costs. There's potential for sodium ion batteries to become cheaper, but even with a 3GWh factory, the numbers aren't that favorable for sodium ion cells. CATL has all the technology to produce them but they're not pursuing it actively. Why? Again, because of the fact they have nearly 500GWh of production capacity invested in lithium ion batteries. It would cannibalize sales and directly compete with what they're currently producing. Sodium ion batteries also suffer from lower energy density, the specific energy is higher than what Xebec mentioned above, the ones I have are around 140-150Wh/kg, but as far as cost vs lead acid, both sodium and lithium batteries are already better than lead acid batteries, especially if you consider their real capacity when factoring deeper discharge cycles. The biggest challenge with sodium ion is the operating voltage range is very very wide, which makes BMS a bigger challenge, a full charge is 4v, low voltage cut-off is as low as 1.5v, a 4 cell pack would need 16v to fully charge and drain down to 6v. It's not difficult to design circuits to handle the wide operating voltage range, but now wire gauge does become a bigger factor since you'd need wires to handle the greater than 2.5x difference in ampacity between when the battery is fully charged and nearly fully discharged. But that's a bit of a tangent with sodium ion cells, let's get back to silicon anodes.

You could argue that silicone anode would be a simple swap for existing lithium ion factories but it's still a huge undertaking. Lithium battery producers are fairly risk averse because if even the smallest defects leave the assembly line, it could cause a catastrophic recall. Major cell companies already invest a ton of money into new chemistries on the horizon, SK even has an affiliate SM Materials making silicon anodes, Samsung is investing heavily into solid state, all these companies are constantly innovating materials with incremental improvements, but don't expect huge leaps in the battery space. With newcomers, without massive financial backing and without a partnering top lithium cell manufacturer to partner up with, it's impossible to make any impact on the battery market. There's dozens of smaller scale players even in China with multi GWh of production capacity and they simply can't compete with the bigger companies in economics of scale, despite the fact they've invested hundreds of millions to billions of dollars.

Besides, at this point, for most part, even current chemistries are good enough, at least for automotive applications. Saving weight would help and improving energy density would as well, but we're not going to get huge improvements in charging time because the massive amounts of electricity needed to quickly charge will start to be limited by the transmission lines, a 100kWh pack with a 10 minute recharge would require 600kW on the transmission line, you'd need medium voltage lines to the charging station to be able to handle even charging one single car. A 250kW Tesla Supercharger works with three phase 480v, that's still 500+ amps, I don't think it's practical to use 480v for chargers higher than 350kW, or at the very least, it would require a LOT of thick wires to distribute from the transformer dropping down medium voltage line to all the individual chargers at a charging station. 1000 MCM cables are almost 2 inches in diameter and can handle 500-800 amps depending on insulation rating, they're not easy to work with, very heavy and very pricey. That'll be a challenge for battery charging architectures as well, we're seeing 800v, 900v and 1000v battery architectures, the wiring needs and insulation on the car itself will become a bigger factor with weight.

Great points, if everyone in a neighborhood even want to charge a small number of cars in limited time, the houses would have to go dark. Excellent point.

 

[count]

On Sodium Ion - I don't think that paticular technology will be impactful unless Lithium gets really expensive (which I don't think will happen). The reason is, yes you save money on one raw material - lithium, but because voltages and energy density is lower, you need more of everything else. More copper, aluminum, seperator material, binder, anode material, and electrolyte, to get the same amount of energy. Not a good trade off unless lithium prices are very high.