Supposedly fusion gets 'interesting' at 10 MK, even though the full temperature you need for fusion is around 50 MK. Basically, if you can get even a few statistically improbably events, you can (a) detect it and (b) dream of setting off a chain reaction. So, ~300 kK * 3x error margin ~= 1 MK, which was what my original assumption was based on. I'll certainly believe they could have gotten that, although I'd still want extremely convincing proof.
The problem is you really need a fluid. It needs to evaporate easily (especially if you don't have an additional source of gas), and it needs to move significantly when a bubble forms. As far as I know, no one has demonstrated cavitation of any kind, let alone sound-induced cavitation, in a solid.
The purpose of the solid is kinda to focus the resonance in the center. It's more just so you get a clean resonance. A sphere has, in theory, exactly one resonance frequency (although if you try this, you'll find many a few kilohertz apart, and some will vanish as temperature changes; this seems to be at least in part due to very small bubbles at the interface between the glass and the water). Any more complex shape has a much more complicated resonance pattern, which means (a, the big deal for humans, less so once it's running) it's harder to find resonance and (b, the point you mentioned, which can be fixed by adding power) less efficiency.
We used "expensive" ($20 or so) piezo drivers with a really clean output in our (slightly ultrasonic) range for our original stuff. Our bigger setup ended up needing more power, so we went to fancier transducers, but they're not needed.
The mic is entirely for your own reading, if you're tuning frequency by hand. Anything that gives a clean output in the 60kHz range is fine, but, of course, this isn't a standard microphone.
Water is warm. it's easy to work with. until there's shown to be a significant difference between solvents, stick with the easy stuff. Everyone agrees that sonoluminescence produces high enough temperatures to disassociate most anything, so all you need is a solvent containing deuterium; pure deuterium is expensive, harder to work with, and not necessarily better.
You absolutely need resonance effects to make sonoluminescence work without having obscene power input.
I was wondering if I should put in a disclaimer that I did my undergrad at the University of Utah... but that's not where I did the sono work, I promise!
It's not hard, really. They'll get drawn there, if you're within a few millimeters, and you can do that by sight.
Much harder is getting the bubble size small enough. These bubbles need to be really small. 7-mil (0.007 inch ID) capillary is a pain in the neck to work with, and coupling it to a whole the same size required creativity. If you want to try, (a) find someone who can drill the whole, you need special equipment. We managed to get a small jewelry bit; (b) get the capillary. It's generally sold in spools of about 100m for some outrageous price ($300 or so), but you can find surplus in lower quantities, and capillary size matters less than hole size. Then, attach them.
Put a thin wire through the whole in the flask. Pour copious amounts of epoxy (not rubber cement, you want something that doesn't absorb too much sound) over the juncture. Thread the capillary on the wire and push until it hits the glass. Let dry, then shave off as much of the epoxy as you can (reduce change in resonance), and pull out the wire. Then realize that the wire broke when you were pulling it out, take a heat gun and melt everything out, thank god you bought more capillary, and try again. Took us four tries for our first rig, only one for our second and third (identical).
In all of the sonoluminesence work I've done, input power has been between 1 and 100 watts. I know people use both lower and higher power, but this is a very reasonable range.
With no additional gas, the bubble size is probably ROUGHLY 10^15 atoms (read as 10^10 - 10^20), depending on a million things. This is at a frequency of roughly (not quite as rough, but close) 10^5 Hz. Assume 10^18 deuterium atoms, for fun, and 0.01% D-D fusion. That gives you (roughly, what, 3.3 MeV for D-D fusion?) around 5kW to play with.
Understand that these numbers are rougher than back of the envelope... these are the kind you do when the envelope will never be found. But if you can pull off fusion at all in sonoluminescence (which is the question at hand), you're pretty much guaranteed decent return on investment.
The stream of bubbles is for demo purposes, you don't want it in a real system. Bubbles are induced in the center by vaporization of the solvent, if no gas is present (actually, it may be that some minimal amount of gas is needed, but there's always/some/, so this is a none-issue).
Keeping the spherical shape, though, it going to either require a container or something close to magic. When you put sound waves through the sphere, it's going to distort, and resonance means that you're in a positive feedback loop. Unless you can apply sound energy equally across the surface of the sphere (from what I understand, applying x-ray energy evenly across a sphere was one of the bigger problems in thermonuclear weapons), a container is needed.
Yep. Less gas gives you higher temperatures, less light. It's a tradeoff. But for show-and-tell, more light is better. It's also much easier to figure out temperature with more light, and then project how temperature increases as the amount of gas gradually decreases; with no extra gas at all, trying to get a reliable spectrum was the most difficult thing I've done, and even then the error bars were huge. (For reference, with no extra gas at all, and degassed water, our original setup, as described, ticked a photomultiplier tube less than a million times a second. That's essentially the number of photons emitted over a significant (1% or so) portion of the sphere. Our next setup was built specifically to make that case more managable, but it was still sketchy.)
This was our original setup, just to make sure our equipment worked; we then went on to bigger and better things, so we could get more reproducability, higher amplitude, pressurized gas on top of the liquid to keep dissolved gas constant and measure change in constituents... all sorts of fun stuff. No video, unfortunately. It wouldn't be really interesting, I don't think. I've always described the appearance as a "star in a jar"... and that's exactly what it looks like. But stars look amazing because there are so many of them, one just sorta sitting alone is just a blue-green point of light, probably too dim to register on most cameras.
Fast neutrons hitting water and acetone give less heinous crap than fast neutrons hitting lead and above, basically.
Thanks for clearing up the solvent... makes sense. Acetone is good stuff for sono, and it has a decent density of hydrogen/deuterium. I'd like to know if they really found an effect they could obtain with acetone and not water...
Also, it's true, we didn't try to recreate the exact same conditions as in this latest paper, mostly because our work predates it; and I don't even know what there solvent is, so I can't even say for sure if we've tested that. But we did reproduce most of the earlier work that lead to the other fifty or sixty claims of 'fusion' in sonoluminescence, with consistent negative results; we also verified the (accepted) fact that solvent doesn't make a huge difference.
If this is true (as mentioned elsewhere, I'm not convinced), it's more than just a method a plasma containment, it's a method of plasma generation. Which, from a sheer elegence perspective (the same one that makes people use Scheme and doubt brane theory) is kinda cool.
Our maximum temperature for sonoluminescence in water was about 280 kK (kilokelvin). Our maximum temperature for sonoluminescence in seeded water (water + hydrogen, for example, although we used water + argon and water + helium; both gave similar results) was around 100 kK. I'll readily believe the second number can improve to approximate the first, but the first just isn't close.
In other substances, nothing seemed quite as good as water. Glycerine and alcohol were both within a factor of two; everything else was lower. Lower molecular density seems to give higher maximum temperature (although I'd have to check the theory to verify this isn't just a coincidence), so trying liquid helium might be cute... but I can't believe it'll help much.
Perhaps I should clarify. We got these results when attempting to reproduce these results, which is why I doubt them. Our results were also consistent with our earlier results trying to estimate the peak temperature possible by sonoluminescence in a given fluid (which is, theoretically, unique for any particular fluid); both results were roughly an order of magnitude smaller than needed for fusion.
Our setup is presumably somewhat different than ours, but here's the summary of the five-minute do-it-your-self sonoluminescence kit:
Take a spherical flask, around 100ml or so. Bigger will mean lower frequencies but higher amplitudes needed. Fill the flask with water from the tap, up until the mensicus is just at the neck of the flask (that is, the water body is as close to spherical as possible). Attach on opposite sides of the flask two speakers, and somewhere else (we just put it between the two speakers, 90 degrees from each, but it doesn't really matter) a microphone.
Hook up a frequency generator to your speakers. Hook up your mic to a 'scope. You'll see the frequency being generated being picked up, slightly muffled and distorted, by the microphone. Tune your frequency until you get resonance; it'll be really, really obvious as the peaks of the mic output become much sharper than the input frequency. The actual frequency depends greatly on the water volume, and is very sensitive to temperature; for our particular setup 48kHz - 52kHz seems about right.
Turn off the light. Allow your eyes about 10 minutes to adjust. With this setup, you'll have light about as bright as a 5th-magnitude star. Any stray light at all will limit your detection. Slowly pump up the amplitude of your input. As the amplitude goes up, resonance frequency changes slightly, so tune as needed. The total amplitude needed is not very high, but it's probably going to be in the top half of a non-amplified signal generator's range.
The gas in the bubble, in this case, is a combination of (some) water vapor and (mostly) outgassed dissolved gasses. That's why we used tap water, above. Bottled water has much less dissolved gasses, so will be much dimmer. Also, water that sits there outgasses, so if you don't change your water it'll get dimmer over time. But we can exploit the fact that it's this added gas that glows, if we want.
Drill a very small hole (seven mil, for us) in the exact bottom of your glass flask. Attach a capilary of the same ID, or a bit more. Attach capilary to a gas canister, and input a low flow rate of gas while running the experiment as above. The idea is to have a near-constant flow of extremely small gas bubbles. If the bubbles are too big, nothing will happen at all; the temperature doesn't get high enough. If there are too many bubbles, you disturb resonance something awful. If the bubbles don't pass through the center, they'll be ignored. But if you get it just right, you'll get a nice burst of light (0th or 1st magnitude) when each bubble goes through, appearing as a constant point of light to the naked eye.
Argon works really nicely for this. Nitrogen works too. You don't want to use anything that dissolves too easily, because it will saturate the water; too much gas outgassing results in bubbles too big to glow. And you'll have to chance the water quite often, because everything will dissolve too much eventually (although helium seems to either dissolve less or just outgas from the top of the flask more quickly).
I presume what they're using in this experiment is hydrogen/deuterium gas, either fed in ordissolved in the water.
Since I should be studying for a midterm, I'll cut off my tutorial now, but feel free to ask more!
I've done a bunch of work in sonoluminescence. It's deeply cool, don't get me wrong. But the highest temperature we were able to measure was about an order of magnitude too low for fusion. Even if our measuring had an error factor of two or three (not impossible, since we had to dope the water to get high enough brightness for using a spectrometer), I'm far from convinced.
That would be... large. My history for the past week (looking in Safari) is well over 20,000 pages... even a year would be a heck of a lot to manage, let alone navigate efficiently. Start UI structures for trees just don't scale that well.
Except there are many more untrue things than true things in the world. If we take it unto ourselves to disprove rigorously everything untrue, we'll never have time for breakfast. Filtering out bozos is a necessary step to making progress in the world.
I find that the categorization I use most often when I meet new people these days is how they respond to 'feeling stupid.' A person can either get frustrated and walk off in a huff, or feel motivated to learn more. I get the sense you are currently in the first category; you will not only be smarter, but probably lead a more fulfilling life, if you manage to move into the second.
Slashdot Mirror
They are, indeed, doing D-D fusion.
Supposedly fusion gets 'interesting' at 10 MK, even though the full temperature you need for fusion is around 50 MK. Basically, if you can get even a few statistically improbably events, you can (a) detect it and (b) dream of setting off a chain reaction. So, ~300 kK * 3x error margin ~= 1 MK, which was what my original assumption was based on. I'll certainly believe they could have gotten that, although I'd still want extremely convincing proof.
The problem is you really need a fluid. It needs to evaporate easily (especially if you don't have an additional source of gas), and it needs to move significantly when a bubble forms. As far as I know, no one has demonstrated cavitation of any kind, let alone sound-induced cavitation, in a solid.
The purpose of the solid is kinda to focus the resonance in the center. It's more just so you get a clean resonance. A sphere has, in theory, exactly one resonance frequency (although if you try this, you'll find many a few kilohertz apart, and some will vanish as temperature changes; this seems to be at least in part due to very small bubbles at the interface between the glass and the water). Any more complex shape has a much more complicated resonance pattern, which means (a, the big deal for humans, less so once it's running) it's harder to find resonance and (b, the point you mentioned, which can be fixed by adding power) less efficiency.
It should be said, also, that our big setup was actually smaller, to concentrate the power a bit more.
We used "expensive" ($20 or so) piezo drivers with a really clean output in our (slightly ultrasonic) range for our original stuff. Our bigger setup ended up needing more power, so we went to fancier transducers, but they're not needed.
The mic is entirely for your own reading, if you're tuning frequency by hand. Anything that gives a clean output in the 60kHz range is fine, but, of course, this isn't a standard microphone.
The luminence of sonoluminescence is definitely black body (heat-generated) radiation, not a discernable spectrum (electrons jumping between states).
Water is warm. it's easy to work with. until there's shown to be a significant difference between solvents, stick with the easy stuff. Everyone agrees that sonoluminescence produces high enough temperatures to disassociate most anything, so all you need is a solvent containing deuterium; pure deuterium is expensive, harder to work with, and not necessarily better.
You absolutely need resonance effects to make sonoluminescence work without having obscene power input.
I was wondering if I should put in a disclaimer that I did my undergrad at the University of Utah... but that's not where I did the sono work, I promise!
It's not hard, really. They'll get drawn there, if you're within a few millimeters, and you can do that by sight.
Much harder is getting the bubble size small enough. These bubbles need to be really small. 7-mil (0.007 inch ID) capillary is a pain in the neck to work with, and coupling it to a whole the same size required creativity. If you want to try, (a) find someone who can drill the whole, you need special equipment. We managed to get a small jewelry bit; (b) get the capillary. It's generally sold in spools of about 100m for some outrageous price ($300 or so), but you can find surplus in lower quantities, and capillary size matters less than hole size. Then, attach them.
Put a thin wire through the whole in the flask. Pour copious amounts of epoxy (not rubber cement, you want something that doesn't absorb too much sound) over the juncture. Thread the capillary on the wire and push until it hits the glass. Let dry, then shave off as much of the epoxy as you can (reduce change in resonance), and pull out the wire. Then realize that the wire broke when you were pulling it out, take a heat gun and melt everything out, thank god you bought more capillary, and try again. Took us four tries for our first rig, only one for our second and third (identical).
In all of the sonoluminesence work I've done, input power has been between 1 and 100 watts. I know people use both lower and higher power, but this is a very reasonable range.
With no additional gas, the bubble size is probably ROUGHLY 10^15 atoms (read as 10^10 - 10^20), depending on a million things. This is at a frequency of roughly (not quite as rough, but close) 10^5 Hz. Assume 10^18 deuterium atoms, for fun, and 0.01% D-D fusion. That gives you (roughly, what, 3.3 MeV for D-D fusion?) around 5kW to play with.
Understand that these numbers are rougher than back of the envelope... these are the kind you do when the envelope will never be found. But if you can pull off fusion at all in sonoluminescence (which is the question at hand), you're pretty much guaranteed decent return on investment.
The stream of bubbles is for demo purposes, you don't want it in a real system. Bubbles are induced in the center by vaporization of the solvent, if no gas is present (actually, it may be that some minimal amount of gas is needed, but there's always /some/, so this is a none-issue).
Keeping the spherical shape, though, it going to either require a container or something close to magic. When you put sound waves through the sphere, it's going to distort, and resonance means that you're in a positive feedback loop. Unless you can apply sound energy equally across the surface of the sphere (from what I understand, applying x-ray energy evenly across a sphere was one of the bigger problems in thermonuclear weapons), a container is needed.
Yep. Less gas gives you higher temperatures, less light. It's a tradeoff. But for show-and-tell, more light is better. It's also much easier to figure out temperature with more light, and then project how temperature increases as the amount of gas gradually decreases; with no extra gas at all, trying to get a reliable spectrum was the most difficult thing I've done, and even then the error bars were huge. (For reference, with no extra gas at all, and degassed water, our original setup, as described, ticked a photomultiplier tube less than a million times a second. That's essentially the number of photons emitted over a significant (1% or so) portion of the sphere. Our next setup was built specifically to make that case more managable, but it was still sketchy.)
This was our original setup, just to make sure our equipment worked; we then went on to bigger and better things, so we could get more reproducability, higher amplitude, pressurized gas on top of the liquid to keep dissolved gas constant and measure change in constituents... all sorts of fun stuff. No video, unfortunately. It wouldn't be really interesting, I don't think. I've always described the appearance as a "star in a jar"... and that's exactly what it looks like. But stars look amazing because there are so many of them, one just sorta sitting alone is just a blue-green point of light, probably too dim to register on most cameras.
Fast neutrons hitting water and acetone give less heinous crap than fast neutrons hitting lead and above, basically.
Thanks for clearing up the solvent... makes sense. Acetone is good stuff for sono, and it has a decent density of hydrogen/deuterium. I'd like to know if they really found an effect they could obtain with acetone and not water...
Please see my other postings in this thread.
Also, it's true, we didn't try to recreate the exact same conditions as in this latest paper, mostly because our work predates it; and I don't even know what there solvent is, so I can't even say for sure if we've tested that. But we did reproduce most of the earlier work that lead to the other fifty or sixty claims of 'fusion' in sonoluminescence, with consistent negative results; we also verified the (accepted) fact that solvent doesn't make a huge difference.
If this is true (as mentioned elsewhere, I'm not convinced), it's more than just a method a plasma containment, it's a method of plasma generation. Which, from a sheer elegence perspective (the same one that makes people use Scheme and doubt brane theory) is kinda cool.
Our maximum temperature for sonoluminescence in water was about 280 kK (kilokelvin). Our maximum temperature for sonoluminescence in seeded water (water + hydrogen, for example, although we used water + argon and water + helium; both gave similar results) was around 100 kK. I'll readily believe the second number can improve to approximate the first, but the first just isn't close.
In other substances, nothing seemed quite as good as water. Glycerine and alcohol were both within a factor of two; everything else was lower. Lower molecular density seems to give higher maximum temperature (although I'd have to check the theory to verify this isn't just a coincidence), so trying liquid helium might be cute... but I can't believe it'll help much.
Perhaps I should clarify. We got these results when attempting to reproduce these results, which is why I doubt them. Our results were also consistent with our earlier results trying to estimate the peak temperature possible by sonoluminescence in a given fluid (which is, theoretically, unique for any particular fluid); both results were roughly an order of magnitude smaller than needed for fusion.
Our setup is presumably somewhat different than ours, but here's the summary of the five-minute do-it-your-self sonoluminescence kit:
Take a spherical flask, around 100ml or so. Bigger will mean lower frequencies but higher amplitudes needed. Fill the flask with water from the tap, up until the mensicus is just at the neck of the flask (that is, the water body is as close to spherical as possible). Attach on opposite sides of the flask two speakers, and somewhere else (we just put it between the two speakers, 90 degrees from each, but it doesn't really matter) a microphone.
Hook up a frequency generator to your speakers. Hook up your mic to a 'scope. You'll see the frequency being generated being picked up, slightly muffled and distorted, by the microphone. Tune your frequency until you get resonance; it'll be really, really obvious as the peaks of the mic output become much sharper than the input frequency. The actual frequency depends greatly on the water volume, and is very sensitive to temperature; for our particular setup 48kHz - 52kHz seems about right.
Turn off the light. Allow your eyes about 10 minutes to adjust. With this setup, you'll have light about as bright as a 5th-magnitude star. Any stray light at all will limit your detection. Slowly pump up the amplitude of your input. As the amplitude goes up, resonance frequency changes slightly, so tune as needed. The total amplitude needed is not very high, but it's probably going to be in the top half of a non-amplified signal generator's range.
The gas in the bubble, in this case, is a combination of (some) water vapor and (mostly) outgassed dissolved gasses. That's why we used tap water, above. Bottled water has much less dissolved gasses, so will be much dimmer. Also, water that sits there outgasses, so if you don't change your water it'll get dimmer over time. But we can exploit the fact that it's this added gas that glows, if we want.
Drill a very small hole (seven mil, for us) in the exact bottom of your glass flask. Attach a capilary of the same ID, or a bit more. Attach capilary to a gas canister, and input a low flow rate of gas while running the experiment as above. The idea is to have a near-constant flow of extremely small gas bubbles. If the bubbles are too big, nothing will happen at all; the temperature doesn't get high enough. If there are too many bubbles, you disturb resonance something awful. If the bubbles don't pass through the center, they'll be ignored. But if you get it just right, you'll get a nice burst of light (0th or 1st magnitude) when each bubble goes through, appearing as a constant point of light to the naked eye.
Argon works really nicely for this. Nitrogen works too. You don't want to use anything that dissolves too easily, because it will saturate the water; too much gas outgassing results in bubbles too big to glow. And you'll have to chance the water quite often, because everything will dissolve too much eventually (although helium seems to either dissolve less or just outgas from the top of the flask more quickly).
I presume what they're using in this experiment is hydrogen/deuterium gas, either fed in ordissolved in the water.
Since I should be studying for a midterm, I'll cut off my tutorial now, but feel free to ask more!
For such a bad joke formula, it's surprising how often it's mildly amusing.
I've done a bunch of work in sonoluminescence. It's deeply cool, don't get me wrong. But the highest temperature we were able to measure was about an order of magnitude too low for fusion. Even if our measuring had an error factor of two or three (not impossible, since we had to dope the water to get high enough brightness for using a spectrometer), I'm far from convinced.
That would be... large. My history for the past week (looking in Safari) is well over 20,000 pages... even a year would be a heck of a lot to manage, let alone navigate efficiently. Start UI structures for trees just don't scale that well.
Games too easy? Play nethack.
Except there are many more untrue things than true things in the world. If we take it unto ourselves to disprove rigorously everything untrue, we'll never have time for breakfast. Filtering out bozos is a necessary step to making progress in the world.
I find that the categorization I use most often when I meet new people these days is how they respond to 'feeling stupid.' A person can either get frustrated and walk off in a huff, or feel motivated to learn more. I get the sense you are currently in the first category; you will not only be smarter, but probably lead a more fulfilling life, if you manage to move into the second.