3 ms·

It's simply early days. People think of this as a fashion industry where one product replaces the new product when it goes out of fashion in the blink of an eye

It's simply early days. People think of this as a fashion industry where one product replaces the new product when it goes out of fashion in the blink of an eye. It simply doesn't work that way with batteries. Scaling up battery production to where it is today has taken decades. These new battery cells need factories that don't exist yet, new machines that are still being designed, they need to benefit from learning effects resulting from volume production that hasn't happened yet, etc. That's not going to happen overnight; it's going to take years/decades. Initial production volume will be low and the resulting products will be very expensive and scarce. There's nothing wrong with the current batteries. They work, they are cheap, they have decent energy density (plenty for essentially all road based use cases), they last thousands of cycles (i.e. 10-20 years at least), and they are being produced by the twh/year. These things aren't going away, they are just going to continue to get better and cheaper for decades to come. It's already a mass production market and the drops in prices is growing the market with new products. Solid state will grab market share very slowly in e.g. high end sports cars, pointlessly large trucks for socker mums and similar snobs (these aren't work vehicles, they are way too expensive for that), electrical planes and drones, and other use cases where price matters less and where energy density matters more. I expect several car companies will make a fair amount of money selling insanely large batteries to people who believe they need to buy those at a big premium. But that's not going to be a mass market any time soon. These will be niche products produced in small numbers and sold at a high price. Mass production won't be happening anytime soon. Most regular people will be driving lithium ion, LFP, sodium ion based cells, or similarly cheap/effective cells for a long time.

Animats 2y ago
The big players are saying small-scale production in 2026-2027, mass production in 2027-2028.[1] The US efforts are tiny startups, but CATL alone has 1,000 people on solid-state battery R&D. Toyota has a big effort.[2] Nissan is building a pilot plant.[3] Samsung is showing prototypes.[4] They're still not here, though. Not even small expensive ones. Military laptops, power tools, and aircraft would use them regardless of cost. When we start seeing some small products with huge energy storage, it's real. [1] https://www.electrive.com/2024/05/30/china-solid-state-battery-research-domestic-companies/ https://www.electrive.com/2024/05/30/china-solid-state-batte... [2] https://carbuzz.com/toyota-solid-state-battery-basic-design-initiated/ https://carbuzz.com/toyota-solid-state-battery-basic-design-... [3] https://global.nissannews.com/en/releases/nissan-shows-in-construction-all-solid-state-battery-pilot-line-in-japan https://global.nissannews.com/en/releases/nissan-shows-in-co... [4] https://www.samsungsdi.com/sdi-now/sdi-news/4041.html https://www.samsungsdi.com/sdi-now/sdi-news/4041.html
- Animats 2y ago
  There's also a question of whether this is cracked by coming up with a battery chemistry and design that's easier to make, or by working through a hard manufacturing process. Examples from the past include EUV semiconductor exposure, and tungsten lamp filaments. In both cases, an insanely hard production process was made to work. With EUV, for most of a decade, process engineers were trying to find some way to avoid the horribly expensive and clunky vaporized-tin EUV light sources. But nothing worked. Not small synchrotrons. A big linear accelerator would work, and SLAC was once borrowed to expose an IC, but that wasn't really feasible. Electron-beams systems remained too slow. So it had to be done the hard way. That's what ASML does. So far, nobody has a better way. Work continues in China on synchrotrons. For tungsten lamp filaments, it was known in the late 1800s that they would work if they could be made in quantity. What's needed is fine wire made of a metal with a very high melting point. Platinum works fine, but costs too much. It took until 1911 to crack the tungsten problem. First, there's a complex sequence of grinding and chemical operations to get tungsten out of ore. There's a smelting step. There's more chemical processing, involving both hydrochloric and hydroflouric acid steps. Not fun. Then washing, separation, and grinding. That just gets you tungsten powder, not wire. So there's a sintering step in a heated press that makes fragile tungsten ingots. Those are annealed, reheated, and rolled down to long round rods. The rods get pointed so they can be started through a die to be drawn into wire. This realigns the crystal structure and the metal becomes ductile. Heating, cooling, and lubrication are involved. After passes through many dies, some made of diamond, fine tungsten wire emerges. General Electric had a big plant in Cleveland doing this for a century, after about two decades of struggling with the process. Then along came LEDs, and the Cleveland twire factory is now a vacant lot. Sometimes, though, the solution is to simplify the thing to be manufactured. A good example was TV camera tubes. The early tubes, the iconocope (Zworklin) and the image dissector (Farnsworth) had poor light sensitivity, but for quite different reasons. The iconoscope integrates light over a full frame time, but lacks a photomultiplier stage to amplify the emitted electrons. The image dissector only counts the light from the moment when the scan beam is hitting the spot being scanned, but it has a photomultiplier to amplify the few electrons knocked loose. So RCA bought the rights to both technologies and combined them into the image orthicon, a tube that cost US$10,000 in the 1950s. An image orthicon integrates light over a full frame, and has a photomultiplier built in, so light sensitivity is good enough to be usable. It was insanely complicated, required a large number of different power supplies and sweep signals, was a pain to adjust, and you needed three of them for color. So a color camera weighed hundreds of pounds and had hundreds of manual adjustments. But it worked and was used by TV stations. Then the vidicon was developed, which is much simpler and can be built with internal color filters. One tube does the whole job. Goodbye, image orthicon. In time, goodbye, RCA. We don't yet know where the solid state battery ends up on this scale. Making a ceramic as a paste on a roll to roll machine is quite a trick. There are some videos of a prototype production line in China which claims to do this. They show lots of battery packaging machinery, which is about the same as it is for lithium-ion, but the ceramic-depositing step isn't shown at all. It probably involves heat and pressure plus other tricks, and they don't want to give those away. There's some worry about making massive investments in a process that rapidly becomes obsolete. There's worry about being run over by someone who does make those investments, takes over the market, and dominates it while someone else is trying to develop the next generation technology. This technology is so valuable that it looks like both approaches are being tried simultaneously. Some will fail.

Animats 2y ago

The big players are saying small-scale production in 2026-2027, mass production in 2027-2028.[1] The US efforts are tiny startups, but CATL alone has 1,000 people on solid-state battery R&D. Toyota has a big effort.[2] Nissan is building a pilot plant.[3] Samsung is showing prototypes.[4] They're still not here, though. Not even small expensive ones. Military laptops, power tools, and aircraft would use them regardless of cost. When we start seeing some small products with huge energy storage, it's real. [1] https://www.electrive.com/2024/05/30/china-solid-state-battery-research-domestic-companies/ https://www.electrive.com/2024/05/30/china-solid-state-batte... [2] https://carbuzz.com/toyota-solid-state-battery-basic-design-initiated/ https://carbuzz.com/toyota-solid-state-battery-basic-design-... [3] https://global.nissannews.com/en/releases/nissan-shows-in-construction-all-solid-state-battery-pilot-line-in-japan https://global.nissannews.com/en/releases/nissan-shows-in-co... [4] https://www.samsungsdi.com/sdi-now/sdi-news/4041.html https://www.samsungsdi.com/sdi-now/sdi-news/4041.html

Animats 2y ago
There's also a question of whether this is cracked by coming up with a battery chemistry and design that's easier to make, or by working through a hard manufacturing process. Examples from the past include EUV semiconductor exposure, and tungsten lamp filaments. In both cases, an insanely hard production process was made to work. With EUV, for most of a decade, process engineers were trying to find some way to avoid the horribly expensive and clunky vaporized-tin EUV light sources. But nothing worked. Not small synchrotrons. A big linear accelerator would work, and SLAC was once borrowed to expose an IC, but that wasn't really feasible. Electron-beams systems remained too slow. So it had to be done the hard way. That's what ASML does. So far, nobody has a better way. Work continues in China on synchrotrons. For tungsten lamp filaments, it was known in the late 1800s that they would work if they could be made in quantity. What's needed is fine wire made of a metal with a very high melting point. Platinum works fine, but costs too much. It took until 1911 to crack the tungsten problem. First, there's a complex sequence of grinding and chemical operations to get tungsten out of ore. There's a smelting step. There's more chemical processing, involving both hydrochloric and hydroflouric acid steps. Not fun. Then washing, separation, and grinding. That just gets you tungsten powder, not wire. So there's a sintering step in a heated press that makes fragile tungsten ingots. Those are annealed, reheated, and rolled down to long round rods. The rods get pointed so they can be started through a die to be drawn into wire. This realigns the crystal structure and the metal becomes ductile. Heating, cooling, and lubrication are involved. After passes through many dies, some made of diamond, fine tungsten wire emerges. General Electric had a big plant in Cleveland doing this for a century, after about two decades of struggling with the process. Then along came LEDs, and the Cleveland twire factory is now a vacant lot. Sometimes, though, the solution is to simplify the thing to be manufactured. A good example was TV camera tubes. The early tubes, the iconocope (Zworklin) and the image dissector (Farnsworth) had poor light sensitivity, but for quite different reasons. The iconoscope integrates light over a full frame time, but lacks a photomultiplier stage to amplify the emitted electrons. The image dissector only counts the light from the moment when the scan beam is hitting the spot being scanned, but it has a photomultiplier to amplify the few electrons knocked loose. So RCA bought the rights to both technologies and combined them into the image orthicon, a tube that cost US$10,000 in the 1950s. An image orthicon integrates light over a full frame, and has a photomultiplier built in, so light sensitivity is good enough to be usable. It was insanely complicated, required a large number of different power supplies and sweep signals, was a pain to adjust, and you needed three of them for color. So a color camera weighed hundreds of pounds and had hundreds of manual adjustments. But it worked and was used by TV stations. Then the vidicon was developed, which is much simpler and can be built with internal color filters. One tube does the whole job. Goodbye, image orthicon. In time, goodbye, RCA. We don't yet know where the solid state battery ends up on this scale. Making a ceramic as a paste on a roll to roll machine is quite a trick. There are some videos of a prototype production line in China which claims to do this. They show lots of battery packaging machinery, which is about the same as it is for lithium-ion, but the ceramic-depositing step isn't shown at all. It probably involves heat and pressure plus other tricks, and they don't want to give those away. There's some worry about making massive investments in a process that rapidly becomes obsolete. There's worry about being run over by someone who does make those investments, takes over the market, and dominates it while someone else is trying to develop the next generation technology. This technology is so valuable that it looks like both approaches are being tried simultaneously. Some will fail.