6 ms·

Most interesting fact from the article. At the system level it took 300 million Joules to produce 3 Million a 1% return - so still a very long way to go! With

Most interesting fact from the article. At the system level it took 300 million Joules to produce 3 Million a 1% return - so still a very long way to go! With this kind of approach I can’t see how fusion can ever be cheap, the capital costs are enormous. Better to focus on standardised production lines for small fission reactors.

Jensson 4y ago
Engineered lasers are over 10x more efficient at emitting light than what they got, and it is very possible they could improve the fusion ignition process to generate more per laser needed, so I wouldn't rule this out yet.
tmikaeld 4y ago
For one, it's theoretically impossible to make lasers 102 times more efficient, especially if you consider the current industrial efficiency of up to 20% (30% under special conditions). For example, the maximum efficiency of a laser is limited by the quantum efficiency of the laser material, which is the ratio of the number of photons emitted by the laser to the number of charge carriers (electrons and holes) injected into the laser material. The quantum efficiency of most laser materials is less than 100%, which means that there is a limit to the efficiency that can be achieved. In addition, the efficiency of a laser is also limited by the thermodynamic laws of thermodynamics, which state that it is not possible to convert heat or any other form of energy into work with 100% efficiency. This means that there is always some amount of waste heat generated when a laser is operated, which limits its efficiency. In other words, fusion is unlikely to be using lasers for the ignitions. UPDATE: People here seem to believe in self-sustained fusion, but this has never been tried or proven to be possible - no one know how much energy is needed for repeated ignition. The current plan is to re-ignite new pellets fed into the containment.
- danbruc 4y ago
  For one, it's theoretically impossible to make lasers 102 times more efficient, especially if you consider the current industrial efficiency of up to 20% (30% under special conditions). That is not true. You can of course not make a 10 % efficient laser more than 10 times more efficient, but there is no limit to how much more efficient you can make a laser in general, you just have to start with an inefficient enough one. And the NIF lasers are quite inefficient, according to Wikipedia - which might have slightly outdated numbers as someone pointed out to me - they turn 422 MJ stored in capacitors into 1.8 MJ UV laser light, that is an efficiency of 0.43 %. Bringing this up to the 20 % you mentioned would already almost be a factor of 50. Also not all of the gain has to come from more efficient lasers, the fusion process could also be made more efficient. But you are of course correct that there is some general limit to laser efficiency, therefore the question is if the fusion gain can compensate this and all additional losses during the conversion into electricity further down the line.
  - TheOtherHobbes 4y ago
    I'm not sure if it's theoretically impossible. But it's certainly an incredible challenge to build a high efficiency laser system that can consistently generate many pulses per second without degrading either the laser head itself or any associated micro-focussing optics. Can we stop pretending this isn't about nuclear weapons research and - basically - continued funding? LLNL is at least a decade behind its initial predictions of ignition. The current lasers may be inefficient but they've been fine-tuned very precisely for optical quality. There's no guarantee more efficient lasers would have the same characteristics and could be focussed in the same way. I'm sure LLNL know the maximum theoretical fusion gain for a realistic pellet design, and it's worth nothing that that number hasn't been mentioned anywhere. Clearly there's only so much energy available for each cycle. If that energy is on the wrong side of what's needed to release it even after efficiency improvements, the entire system isn't workable, no matter how it's re-engineered.
    - danbruc 4y ago
      My point was purely about the claimed impossibility of a hundredfold efficiency increase which is only true if you are already about 1 % efficient. I have no real opinion on what the NIF is doing or not, at best my layman judgment would be that laser fusion is way further from becoming practical than tokamaks or stellerators.
      - tmikaeld 4y ago
        You're right that we don't have enough information here, I've been reading a lot about it's theory and practicality throughout the years. There's also theories about ignition in parallell, by using multiple pellets. Since the first step has been taken, there's a lot of avenues for improvement. However, one also has to take into account the "chaos" created by the ignition in the first place, re-introducing an additional pellet or chain reaction has it's own challenges.
    - scythe 4y ago
      >Can we stop pretending this isn't about nuclear weapons research The laser practicality issue that prevents this from directly becoming a power source would also be a major barrier to its application as a weapon. The laser fires in the UV-B (351 nm), which is scattered and attenuated by air, to say nothing of smoke or dust; it also requires incredibly high targeting precision (<2 mm target diameter) and consequently precise placement of a target weighing only milligrams. Additionally, the various optical components of the laser must be very well aligned, which is difficult to achieve in any battlefield conditions. And the whole thing is a very obvious and vulnerable target. I ascribe a small possibility to its utility as a weapon in space, but practically zero on Earth without other major developments. >I'm sure LLNL know the maximum theoretical fusion gain for a realistic pellet design, and it's worth nothing that that number hasn't been mentioned anywhere. I wouldn't be so sure. Fusion is in general quantum chromodynamics, which is not so well characterized (being the subject of the famous YM mass gap conjecture). Even in this case it was stated that the yield exceeded expectations and damaged the sensors, which was probably not desired.
skelpmargyar 4y ago
It's better to focus on standardised production lines for small fission reactors currently if your goal is to produce energy for the power grid right now, but that's not what cutting-edge research or science is about. Whoever launched the first firework probably didn't see how people could walk on the moon either. Science is rarely a linear path. It's also rarely "I have this problem so I solved it this way." That's more engineering than science. It's really disappointing to me to see people, for lack of a better term, just shit on this accomplishment on HN when it's some seriously amazing stuff. It seems like a lot of people here don't understand science, they only understand engineering and then think about engineering mostly from a dull business and product oriented perspective.
- nayroclade 4y ago
  Except it isn't the HN readership who choose to constantly frame these stories in terms of engineering. From the article: "These and other scientific, technological and engineering hurdles will need to be overcome before fusion will produce electricity for your home. Work will also need to be done to bring the cost of a fusion power plant well down from the US$3.5 billion of the National Ignition Facility. These steps will require significant investment from both the federal government and private industry." If fusion research scientists continue to insist that their research has any viable path to use in commercial power generation, and to demand large amounts on funding on that basis, then they should expect to be critiqued on that same basis.
  - skelpmargyar 4y ago
    What that that paragraph means is that things that do not exist will cost money to bring into existence. A significant amount. It's not a critique. This is not new. This is not a fusion problem. There is an initial cost and development to any new technology. This technology is particularly difficult, however. It's also sustainable real renewable energy. You have to worry about actually being able to invent something before you can worry about saving money on that thing. Look up how much the Manhattan Project cost! This shit is not software engineering; It's really really fucking complicated it will cost a load of money. We don't really know what path fusion research will lead to. It's science. We don't know what we could find out tomorrow that could apply this research. But even the promises of commercial power generation alone should be enough to keep funding the project whether it will happen in 50 years or 100 years. It's not like they aren't making progress. You can't rush research and also under fund them.
throw827474737 4y ago
Depends on what you want to achieve, but with that attitude humanity would not have gotten anywhere.
noobermin 4y ago
People can do multiple things at once...or rather, different people can do different things simultaneously, there is no need to "better focus on" x over y.
ThePhysicist 4y ago
Well, early transistors were 3-5 orders of magnitude larger, slower and more dissipative than those we have today. So by the same argument, the computer revolution would have seemed impossible just 50 years ago.
- fragmede 4y ago
  Transistors? Computers before them used vacuum tubes. In 1943 Thomas Watson, chairman of IBM, said, "I think there's a world market for maybe five computers." Now, it would be easy for us to laugh at him, thanks to us being almost 80 years in the future, but on the topic of "what does the future hold for us", put yourself in Watson's place. You're reasonably smart and knowledgeable, and understand what goes into a computer, and thus you'd make the same claim. Fusion is similarly incomprehensible to us here and now. There are untold advancements in materials sciences and engineering technologies that need to happen before it's possible so we have to invest in trying to so we can make those advancements in order to make it possible.
  - pavlov 4y ago
    Just 75 years after Thomas Watson said that, we were already doing EUV lithography, a process whose description and commercial incentives would have been totally incomprehensible to him: "In our laser-produced plasma (LPP) source, molten tin droplets of around 25 microns in diameter are ejected from a generator at 70 meters per second. As they fall, the droplets are hit first by a low-intensity laser pulse that flattens them into a pancake shape. Then a more powerful laser pulse vaporizes the flattened droplet to create a plasma that emits EUV light. To produce enough light to manufacture microchips, this process is repeated 50,000 times every second." https://www.asml.com/en/technology/lithography-principles/light-and-lasers https://www.asml.com/en/technology/lithography-principles/li... Transistors, microchips and lasers were not invented yet in 1943. How would one even guess that it could soon be a viable business to build a giant machine that shoots molten tin droplets at 50,000 Hz to produce a particular bandwidth of light so that you could create billions of tiny computers out of sand? Hopefully fusion is on a similar path where the description of 2100's commercial reactor will sound similarly incomprehensible to us.
  - 4y ago
mirekrusin 4y ago
Ignition doesn't have to be net positive. It just needs to enter into state of self sustained reaction that is net positive. Ignition itself can be horribly inefficient. Nobody is designing fussion where ignition's output is directly used for the whole system to reignite again. It would be equivalent to moving hand with match that fires a match, where energy from ignition is used to trigger another hand to move another match.
- hoseja 4y ago
  This isn't a system that can ever do a sustained reaction.
  - world-set-free 4y ago
    not yet
subradios 4y ago
Energy production has always been an enormously capital intensive venture. R&D labs aren't known for their industrial optimization unless that's their research goal. They proved this from the perspective of the fuel pellet - now they have to increase energy yield (larger fuel?) , reduce total system power, etc. That said, they doubled energy yield in less than a year. The early exponential function on emerging technologies is always fun.
chongli 4y ago
That's not the only big challenge. The fuel pellets are 1mm diameter hollow spheres machined out of diamond to extreme precision and filled with a tiny amount of deuterium-tritium gas inside. The pellets need to be carefully cooled to cryogenic temperatures so that the gas condenses evenly around the inside of the pellet, in order to maintain the near-perfect geometry needed for ignition. Once the pellet is hit with the x-rays produced by the lasers striking the interior of the chamber (hohlraum), it compresses and the fuel undergoes fusion, vaporizing the whole package. So it takes an incredible amount of precision engineering to produce these tiny diamond pellets of fuel which then produce an even tinier amount of energy while being destroyed in the process. There is no indication whatsoever that the pellets can be scaled up due to the incredible difficulty of scaling the precision geometry involved. So I remain incredibly skeptical that this is anything more than hype for a project that's really about maintaining nuclear weapons stockpiles. I sincerely doubt this approach will lead to a real production power plant without some major research into how to overcome the requirement for extreme precision geometry and a similar effort to scale up the size of the reaction. Then throw the laser efficiency issue into the mix!
- generalizations 4y ago
  One of the things about reaching this point is that we can reconsider just how much of that complex process is required - like getting the poc of some code working, then working backwards to strip away what isn't necessary.
- chasil 4y ago
  The hohlraum cylinder itself is also made of gold. Could it be made from a metal that is in greater supply?
- hinkley 4y ago
  Do you think they’ll drop the cubes into the reactor, The Expanse style, or have to move them in from outside between firings, while also keeping them cryogenically cooled? That sounds like a slow process, especially in a chamber that just had an artificial sun inside it and is running a power generator off of the heat produced. Could they really fire this design more than twice a minute?
xbmcuser 4y ago
Solar panel and power cost were 100s of times more than they are today. Costs will come down that is actually not the problem having a sustainable fusion reaction is. How would you know how efficiently you can build something when you don't know what to build in the first place. If we can get sustainable fusion reaction then only should peope worry about how to bring down the costs.
- moloch-hai 4y ago
  Even if heat from fusion could be extracted at exactly zero cost, thermal fusion would still be uncompetitive. There is no way around it, wish as you might. Aneutronic fusion could possibly have a future, but hardly anybody is working on that.
nottorp 4y ago
This is still fundamental research not commercial engineering. Like everything good before it. Even in engineering problems, two people started a little electric car company in 2003. No, youknowwho wasn't involved yet. Would you have said "i can't see how electric cars can ever be workable" then?
graycat 4y ago
Hmm .... Mars has nearly no atmosphere, and, thus, nearly no bad weather, clouds, fog, rain, etc. That is, it's sunshine all the time! Sooo, for Mars, just put the solar panels near the equator and fairly densely all around the circumference. Then, wow, have solar power 100% of the time! Do the same on Earth??? Ah, have some jungles, mountains, two major oceans, and a lot of bad weather. For the oceans, sure, have floating solar arrays. Right, need to think about how deep the power cables to the shore would be, 3-7 miles down and then back up? But the Internet has some cables across oceans!! Sooo, we could also have power cables??? Ah, sounds like only for Mars!

Jensson 4y ago

Engineered lasers are over 10x more efficient at emitting light than what they got, and it is very possible they could improve the fusion ignition process to generate more per laser needed, so I wouldn't rule this out yet.

tmikaeld 4y ago

For one, it's theoretically impossible to make lasers 102 times more efficient, especially if you consider the current industrial efficiency of up to 20% (30% under special conditions). For example, the maximum efficiency of a laser is limited by the quantum efficiency of the laser material, which is the ratio of the number of photons emitted by the laser to the number of charge carriers (electrons and holes) injected into the laser material. The quantum efficiency of most laser materials is less than 100%, which means that there is a limit to the efficiency that can be achieved. In addition, the efficiency of a laser is also limited by the thermodynamic laws of thermodynamics, which state that it is not possible to convert heat or any other form of energy into work with 100% efficiency. This means that there is always some amount of waste heat generated when a laser is operated, which limits its efficiency. In other words, fusion is unlikely to be using lasers for the ignitions. UPDATE: People here seem to believe in self-sustained fusion, but this has never been tried or proven to be possible - no one know how much energy is needed for repeated ignition. The current plan is to re-ignite new pellets fed into the containment.

danbruc 4y ago
For one, it's theoretically impossible to make lasers 102 times more efficient, especially if you consider the current industrial efficiency of up to 20% (30% under special conditions). That is not true. You can of course not make a 10 % efficient laser more than 10 times more efficient, but there is no limit to how much more efficient you can make a laser in general, you just have to start with an inefficient enough one. And the NIF lasers are quite inefficient, according to Wikipedia - which might have slightly outdated numbers as someone pointed out to me - they turn 422 MJ stored in capacitors into 1.8 MJ UV laser light, that is an efficiency of 0.43 %. Bringing this up to the 20 % you mentioned would already almost be a factor of 50. Also not all of the gain has to come from more efficient lasers, the fusion process could also be made more efficient. But you are of course correct that there is some general limit to laser efficiency, therefore the question is if the fusion gain can compensate this and all additional losses during the conversion into electricity further down the line.
- TheOtherHobbes 4y ago
  I'm not sure if it's theoretically impossible. But it's certainly an incredible challenge to build a high efficiency laser system that can consistently generate many pulses per second without degrading either the laser head itself or any associated micro-focussing optics. Can we stop pretending this isn't about nuclear weapons research and - basically - continued funding? LLNL is at least a decade behind its initial predictions of ignition. The current lasers may be inefficient but they've been fine-tuned very precisely for optical quality. There's no guarantee more efficient lasers would have the same characteristics and could be focussed in the same way. I'm sure LLNL know the maximum theoretical fusion gain for a realistic pellet design, and it's worth nothing that that number hasn't been mentioned anywhere. Clearly there's only so much energy available for each cycle. If that energy is on the wrong side of what's needed to release it even after efficiency improvements, the entire system isn't workable, no matter how it's re-engineered.
  - danbruc 4y ago
    My point was purely about the claimed impossibility of a hundredfold efficiency increase which is only true if you are already about 1 % efficient. I have no real opinion on what the NIF is doing or not, at best my layman judgment would be that laser fusion is way further from becoming practical than tokamaks or stellerators.
    - tmikaeld 4y ago
      You're right that we don't have enough information here, I've been reading a lot about it's theory and practicality throughout the years. There's also theories about ignition in parallell, by using multiple pellets. Since the first step has been taken, there's a lot of avenues for improvement. However, one also has to take into account the "chaos" created by the ignition in the first place, re-introducing an additional pellet or chain reaction has it's own challenges.
  - scythe 4y ago
    >Can we stop pretending this isn't about nuclear weapons research The laser practicality issue that prevents this from directly becoming a power source would also be a major barrier to its application as a weapon. The laser fires in the UV-B (351 nm), which is scattered and attenuated by air, to say nothing of smoke or dust; it also requires incredibly high targeting precision (<2 mm target diameter) and consequently precise placement of a target weighing only milligrams. Additionally, the various optical components of the laser must be very well aligned, which is difficult to achieve in any battlefield conditions. And the whole thing is a very obvious and vulnerable target. I ascribe a small possibility to its utility as a weapon in space, but practically zero on Earth without other major developments. >I'm sure LLNL know the maximum theoretical fusion gain for a realistic pellet design, and it's worth nothing that that number hasn't been mentioned anywhere. I wouldn't be so sure. Fusion is in general quantum chromodynamics, which is not so well characterized (being the subject of the famous YM mass gap conjecture). Even in this case it was stated that the yield exceeded expectations and damaged the sensors, which was probably not desired.

skelpmargyar 4y ago

It's better to focus on standardised production lines for small fission reactors currently if your goal is to produce energy for the power grid right now, but that's not what cutting-edge research or science is about. Whoever launched the first firework probably didn't see how people could walk on the moon either. Science is rarely a linear path. It's also rarely "I have this problem so I solved it this way." That's more engineering than science. It's really disappointing to me to see people, for lack of a better term, just shit on this accomplishment on HN when it's some seriously amazing stuff. It seems like a lot of people here don't understand science, they only understand engineering and then think about engineering mostly from a dull business and product oriented perspective.

nayroclade 4y ago
Except it isn't the HN readership who choose to constantly frame these stories in terms of engineering. From the article: "These and other scientific, technological and engineering hurdles will need to be overcome before fusion will produce electricity for your home. Work will also need to be done to bring the cost of a fusion power plant well down from the US$3.5 billion of the National Ignition Facility. These steps will require significant investment from both the federal government and private industry." If fusion research scientists continue to insist that their research has any viable path to use in commercial power generation, and to demand large amounts on funding on that basis, then they should expect to be critiqued on that same basis.
- skelpmargyar 4y ago
  What that that paragraph means is that things that do not exist will cost money to bring into existence. A significant amount. It's not a critique. This is not new. This is not a fusion problem. There is an initial cost and development to any new technology. This technology is particularly difficult, however. It's also sustainable real renewable energy. You have to worry about actually being able to invent something before you can worry about saving money on that thing. Look up how much the Manhattan Project cost! This shit is not software engineering; It's really really fucking complicated it will cost a load of money. We don't really know what path fusion research will lead to. It's science. We don't know what we could find out tomorrow that could apply this research. But even the promises of commercial power generation alone should be enough to keep funding the project whether it will happen in 50 years or 100 years. It's not like they aren't making progress. You can't rush research and also under fund them.

throw827474737 4y ago

Depends on what you want to achieve, but with that attitude humanity would not have gotten anywhere.

noobermin 4y ago

People can do multiple things at once...or rather, different people can do different things simultaneously, there is no need to "better focus on" x over y.

ThePhysicist 4y ago

Well, early transistors were 3-5 orders of magnitude larger, slower and more dissipative than those we have today. So by the same argument, the computer revolution would have seemed impossible just 50 years ago.

fragmede 4y ago
Transistors? Computers before them used vacuum tubes. In 1943 Thomas Watson, chairman of IBM, said, "I think there's a world market for maybe five computers." Now, it would be easy for us to laugh at him, thanks to us being almost 80 years in the future, but on the topic of "what does the future hold for us", put yourself in Watson's place. You're reasonably smart and knowledgeable, and understand what goes into a computer, and thus you'd make the same claim. Fusion is similarly incomprehensible to us here and now. There are untold advancements in materials sciences and engineering technologies that need to happen before it's possible so we have to invest in trying to so we can make those advancements in order to make it possible.
- pavlov 4y ago
  Just 75 years after Thomas Watson said that, we were already doing EUV lithography, a process whose description and commercial incentives would have been totally incomprehensible to him: "In our laser-produced plasma (LPP) source, molten tin droplets of around 25 microns in diameter are ejected from a generator at 70 meters per second. As they fall, the droplets are hit first by a low-intensity laser pulse that flattens them into a pancake shape. Then a more powerful laser pulse vaporizes the flattened droplet to create a plasma that emits EUV light. To produce enough light to manufacture microchips, this process is repeated 50,000 times every second." https://www.asml.com/en/technology/lithography-principles/light-and-lasers https://www.asml.com/en/technology/lithography-principles/li... Transistors, microchips and lasers were not invented yet in 1943. How would one even guess that it could soon be a viable business to build a giant machine that shoots molten tin droplets at 50,000 Hz to produce a particular bandwidth of light so that you could create billions of tiny computers out of sand? Hopefully fusion is on a similar path where the description of 2100's commercial reactor will sound similarly incomprehensible to us.
- 4y ago

mirekrusin 4y ago

Ignition doesn't have to be net positive. It just needs to enter into state of self sustained reaction that is net positive. Ignition itself can be horribly inefficient. Nobody is designing fussion where ignition's output is directly used for the whole system to reignite again. It would be equivalent to moving hand with match that fires a match, where energy from ignition is used to trigger another hand to move another match.

hoseja 4y ago
This isn't a system that can ever do a sustained reaction.
- world-set-free 4y ago
  not yet

subradios 4y ago

Energy production has always been an enormously capital intensive venture. R&D labs aren't known for their industrial optimization unless that's their research goal. They proved this from the perspective of the fuel pellet - now they have to increase energy yield (larger fuel?) , reduce total system power, etc. That said, they doubled energy yield in less than a year. The early exponential function on emerging technologies is always fun.

chongli 4y ago

That's not the only big challenge. The fuel pellets are 1mm diameter hollow spheres machined out of diamond to extreme precision and filled with a tiny amount of deuterium-tritium gas inside. The pellets need to be carefully cooled to cryogenic temperatures so that the gas condenses evenly around the inside of the pellet, in order to maintain the near-perfect geometry needed for ignition. Once the pellet is hit with the x-rays produced by the lasers striking the interior of the chamber (hohlraum), it compresses and the fuel undergoes fusion, vaporizing the whole package. So it takes an incredible amount of precision engineering to produce these tiny diamond pellets of fuel which then produce an even tinier amount of energy while being destroyed in the process. There is no indication whatsoever that the pellets can be scaled up due to the incredible difficulty of scaling the precision geometry involved. So I remain incredibly skeptical that this is anything more than hype for a project that's really about maintaining nuclear weapons stockpiles. I sincerely doubt this approach will lead to a real production power plant without some major research into how to overcome the requirement for extreme precision geometry and a similar effort to scale up the size of the reaction. Then throw the laser efficiency issue into the mix!

generalizations 4y ago
One of the things about reaching this point is that we can reconsider just how much of that complex process is required - like getting the poc of some code working, then working backwards to strip away what isn't necessary.
chasil 4y ago
The hohlraum cylinder itself is also made of gold. Could it be made from a metal that is in greater supply?
hinkley 4y ago
Do you think they’ll drop the cubes into the reactor, The Expanse style, or have to move them in from outside between firings, while also keeping them cryogenically cooled? That sounds like a slow process, especially in a chamber that just had an artificial sun inside it and is running a power generator off of the heat produced. Could they really fire this design more than twice a minute?

xbmcuser 4y ago

Solar panel and power cost were 100s of times more than they are today. Costs will come down that is actually not the problem having a sustainable fusion reaction is. How would you know how efficiently you can build something when you don't know what to build in the first place. If we can get sustainable fusion reaction then only should peope worry about how to bring down the costs.

moloch-hai 4y ago
Even if heat from fusion could be extracted at exactly zero cost, thermal fusion would still be uncompetitive. There is no way around it, wish as you might. Aneutronic fusion could possibly have a future, but hardly anybody is working on that.

nottorp 4y ago

This is still fundamental research not commercial engineering. Like everything good before it. Even in engineering problems, two people started a little electric car company in 2003. No, youknowwho wasn't involved yet. Would you have said "i can't see how electric cars can ever be workable" then?

graycat 4y ago

Hmm .... Mars has nearly no atmosphere, and, thus, nearly no bad weather, clouds, fog, rain, etc. That is, it's sunshine all the time! Sooo, for Mars, just put the solar panels near the equator and fairly densely all around the circumference. Then, wow, have solar power 100% of the time! Do the same on Earth??? Ah, have some jungles, mountains, two major oceans, and a lot of bad weather. For the oceans, sure, have floating solar arrays. Right, need to think about how deep the power cables to the shore would be, 3-7 miles down and then back up? But the Internet has some cables across oceans!! Sooo, we could also have power cables??? Ah, sounds like only for Mars!