Light from Sonic Implosions
By Alex Wurden
Grade 8, Los Alamos Middle School
Feb. 10, 1996
In this experiment I made sonoluminescence (SL) by focusing sound waves in water, to levitate a single bubble in the middle of a flask. The sound was higher frequency than you can hear, at 28.2 kilohertz. SL occurred in cold, degassed water, at 8-35 watts power level. When I added 5-10% glycerin, it made the SL light ten times brighter. I used my photo-multiplier tube to measure on an oscilloscope the intensity of the light and duration of the flash, which is less than 1/100 thousandths of a second. I also noticed that the phase of the light flash in relation to the sound wave can be changed by altering power, temperature, and the position of the bubble inside the flask. Too much power would smash the bubble, and you would have to inject a new bubble. By looking in a very dark room, I could see individual photon blips on the scope, and so I could estimate that 50,000 photons come out of SL in each burst of light. There are 1.5 billion photons produced per second! With an enlarged picture, I calculated that the light emission from the bubble is less than 250 microns wide on average. Finally, I mapped out the resonances in the general area of 28.2 kHz for current, voltage, and microphone peaks.
Introduction
In this experiment I made light from sonic implosions, known as sonoluminescence. I did this with two piezo-electric discs and a microphone, a 100 ml flask, an Oscilloscope, Audio Amp, and Frequency Generator.
Sonoluminescence occurs when a small bubble of air is trapped in the middle of water with strong, super high sound waves that resonate with the container they're in. Under the right conditions, the bubble will collapse when the sound waves hit it, and create a mini-shock wave that crushes the air down into a super dense point. Then the bubble recoils and shoots back out to regular size. It does this 28,200 times a second! While it is in this very compressed, dense state, it gets hot. Some scientists estimate that it gets between 5000 degrees C, to much hotter than surface of the sun. Theoretically, this is why it gives off the blue, and largely UV spectrum light. This phenomenon is being investigated because it is thought that it can start the fusion process, instead of using costly, super-powerful lasers focused on a tiny spec. It is also being investigated because cavitations like this is corroding Navy propellers, and bubble cavitation is making submarines noisy.
A lot of complex factors are involved in getting the conditions right for Sonoluminescence (SL). Among them is matching the resonance frequency to the electric frequency by adjusting inductance and magnetic fields. Another is properly degassing all of the air dissolved in the water I need to use. Finding a power level on the Audio Amplifier to make the bubble stable, and to keep it from dissolving proved to be my main difficulty in this experiment.
Questions
I want to try to make light from sonic implosions, which I read about in Scientific American. The phenomenon is called sonoluminescence. I want to study it in depth, to find details like brightness, size, when it occurs, and how it happens. I also want to know how all the equipment I'm going to need works.
Hypothesis
Table 1: Materials used in my project
Audio Amplifier LOTS of wire, all kinds Rubber Stoppers Oscilloscope Vacuum Pump Desklight Frequency Generator Multipurpose Digital Meter Kodak 1000 Film and Camera 2 Piezo-Driver Discs Photo-multiplier Tube Solder and Soldering Iron Piezo-Microphone Lenses Black Tape Coaxial Cable Toilet Paper Rolls Cardboard Scraps 80-200 ml Florence Flasks Inductors LED for Lens Focus testing 1000 ml Erinmeyer Capacitors Glycerin Boiling Flask Filter Paper Resistors and inductors Distilled or tap water
I borrowed a Heathkit oscilloscope from a friend, and eventually used a LANL rough pump, and LANL frequency generator (after my own generator stopped working). I used the photo-multiplier detector setup that I built from my last year's science fair project
Assembly
I first took my two driving transducers, and on one side of them soldered three thin wires. Three of them makes a tripod to rest against the glass, and gives me a spare in case one breaks. On the opposite side I only put one wire. Thin wire is best because it minimizes sound loss. Then I glued the two discs at opposite sides of the flask as accurately as possible, so the sound waves inside would collide equally and make a more precise, stable bubble. On the bottom I put the third disc, the microphone. It had only one wire on each side, just because it was so tiny. For gluing, I used a hard, five-minute epoxy that comes in two separate bottles. I used this because it provided a solid material for the sound to travel through, and it chips off the glass easily for cleaning or realignment.
Now that the flask was ready to go, I soldered my driver wires into a circuit containing resistors, inductors, monitor points for the Oscilloscope and Multi-Meter, the Audio Amp, and the Frequency Generator.
The main problem in all the circuitry above is getting the inductor tuned to the resonant frequency of the jar. Also is the very fragile wire coming from the discs which breaks, and the inductors. One of the inductors is very small, and when the power on the Amp is at its maximum for sonoluminescence, (SL) it overheats and limits operating time. I also had to make a home made inductor using five spools of thick Radio Shack wire. It has over 250 feet of wire on it for a mere 4 mH of inductance, and took a while to wind.
At first I used a 250 ml flask for the jar, but I quickly found out that its resonant frequency was too low (about 20 kHz), and I could hear it! This is bad, because it is a very annoying sound. By switching to a smaller 100 ml flask, the resonant frequency is higher, and you cannot hear most of the power, which is at 28 kHz. This was better anyway, because the inductor can be smaller. The inductor is used to balance out the capacitance of the piezo-electric disks, in the form of an LC circuit.
Click here for Graph of resonances
Figure 1: The electrical resonance due to the L-C circuit must be at the same frequency as the acoustic resonance of the glass container.
The electrical resonance is set by Click here for Equation , but the acoustic resonance of the jar depends on the diameter of the jar and the speed of sound in water (v=1500 meters/second), approximately as 1.1 x v/d. The flask is 6.8 cm in diameter, and resonates at 27.5 to 28.4 kHz, depending somewhat on the water level. The capacitance C is 1.2 nanoFarads, and my inductor is about 25.5 mH. I wound the inductor by trial and error, in order to get the electrical and acoustic resonance to match. I made a plot of the current, voltage, and acoustic pickup as a function of frequency near the resonance. The acoustic signal shows several very narrow peaks. My dad says this is probably due to the thickness of the glass.
For my Photo-Multiplier detector tube, I needed to collect as much light from
the SL as possible. To do this, I read some optics books, and made a simple
relay telescope. I tested different lenses, and found their focal lengths.
After sliding them around and projecting the image of a LED on white paper
clearly, I marked them down on cardboard and built mountings. Then I cut
toilet paper tubes of different sizes to fit in-between the components to act
as shielding against background light.
Procedure
Now that the basic apparatus was built, I needed to tune it, and get some actual water ready. According to the articles I read, SL requires water that has been carefully degassed, in order for it to work. First I took my 1000 ml flask, a funnel, and filter paper, and then dripped in water until it is just full enough so that it won't boil over. I put a rubber stopper with a pipe and hose on it securely over the top and start heating, until the water won't boil any faster. Then I take it off the heat, and as soon as the bubbles stop coming up, I clamp off the hose as securely as I can, and let it sit for a while. As soon as it's cool enough, I put it in the refrigerator, to make it even colder. As you can see, this is going to create a super strong suction when the steam condenses, and will draw any gas dissolved in the water out that hadn't been driven off by boiling. I need a really sturdy flask with no cracks, preferably Pyrex. If you can get a real vacuum pump, it works even better for degassing, because it makes a vacuum above the water, and pulls off the air (actually boils the water at room temperature!) and does a more thorough job.
Now I scrub out the 100 ml flask and pour in the freshly degassed water. After I make sure the wires aren't touching and there are no short circuits, I power up my instruments.
Each time before use, the Inductance must be calibrated by moving the coils closer together or farther apart. This is to get the voltage and current in phase, which maximizes the electric current in the circuit, and is the most efficient way to make sound waves in the jar.
I take a straw, put it in the liquid, and take it out with my thumb on the
top, to plop some water. This creates enough of a disturbance so that if I
have the right frequency and power levels, a bubble will be trapped in the
middle. However, it is usually unstable one way or another, so that it shrinks
away into nothingness or it gets so big and wiggly that it can overcome the
pressure of the sound and float to the top. I keep repeating this procedure
and adjusting my frequency and power until the bubble is relatively small and
perfectly stable, after the initial unstable forming period for about five
seconds, of course. Then I turn out the lights, and look very carefully. It
has to be totally dark to be able to see it though, and so all the indicator
lights have to covered up with black tape or paper.
Observations
· My first observation in this experiment was little sqiggly lines on the acoustic trace of the oscilloscope when you crank up the power on the drive circuit. I found these distortions occurred when the bubble grew to extremely large and unstable proportions (usually). These squiggles are called harmonics. They are because the sound frequency is changed as the driving force vibrates the large bubble. Later on I found that harmonics usually accompany Sonoluminescence.
· I noticed that the meniscus of the water level in the flask changes greatly when power is set very high, and is tweaked a little bit. The meniscus can often double its curvature, so it is a big difference, and it shows that the sound can really affect the water.
· The addition of as much Glycerin as the water can dissolve, about 20%, makes sonoluminescence incredibly easier to make. If SL is made without it, and then Gylercin is added, the Glycerin makes the bubble became about five to ten times brighter.
· I took pictures of the sonoluminescence, using fast ASA 1000 Kodak film,
an f 3.6 100 mm closeup lens, with from 1 second to 40 second exposures. One of
the pictures is shown below:
Click here for Picture of Sonoluminescence
Figure 2: Time exposure of the sonoluminescence light in the 100 ml
flask.
· Power affects bubble stability. If the power is too low, there will not be enough force to hold the bubble in the middle, and it will float to the top. If the power is too high, it will either crush the bubble, or excite the water around it so much that it (the bubble) grows rapidly and floats to the top anyway. SL takes place at the very verge of the high power limit which will crush the bubble. I got SL to work between 8 and 35 watts (as measured on the stereo amp meter, but actually only 1.5 to 6.5 watts from the current and voltage measured in the circuit.
· If the water is too degassed, then the bubble made by the plop will re-dissolve itself into the water and shrink away in less than five seconds. If the water is not degassed enough, the bubble will take on air, and float against the pressure to the top. SL somehow manages to occur between these two limits.
· When I switched from boiling and condensing water to degas it, to a vacuum pump, it worked a lot better. After about five seconds of pumping, the water suddenly turns pure white with fizz coming up. That only lasts a few seconds though. It settles down to a level comparable to Sprite. Three minutes or so of that, and it looks like it is boiling, but it isn't. This gradually recedes into total calm. This is when it is done. I have to take the pump off at this point or it will start sucking the water. When I'm degassing with the pump, it is a good idea to fill it very shallow with water, so there is more surface area for the pump to act on, and so that the water in the bottom can still be affected without having to shake the flask from time to time.
· The two most important tricks are to have carefully degassed water, and
to do a power scan after putting a bubble into the jar, in order to slowly
raise the power on a bubble. I used the light detector and the oscilloscope to
fine tune the frequency, power, and detector position, in order to get the
brightest output.
Conclusions
I think that this experiment was quite successful. It is a very complicated system. On my second try, I made Sonoluminescence for a few seconds, from six or seven bubble attempts, but it wouldn't last. However, it then took me about a dozen tries over a two-month period before I could get it to light up again. Finally, as I improved my techniques, I made Sonoluminescence many times, and once it stayed continuously for twenty minutes! It looks like a small blue spark or very tiny star. It is extremely bright for its size. The easiest way to see it by eye, is to look away from it. The best way to make it is to fine tune your inductance, add the Glycerin, find a resonance frequency, and do a slow power scan on the Audio Amp. The power scan, going from low to high is crucial because it has to be within about half a watt to make SL. The other conditions (such as the amount of liquid in the jar) aren't as critical.
I think it is practically impossible to measure the temperature inside the bubble because if you stick a thermometer in, it would either break (the temperature is estimated to be between 5000 C to much hotter than the sun) or disrupt the sound-waves making the SL. Spectroscopy is difficult because the main part of the light emitted is in the UV zone, and ultra-violet light can neither travel through glass nor water. Water, plain or with glycerin, is almost the only substance that SL has been made in so far (this is not understood yet....maybe because of the viscosity, and ability to dissolve lots of different gases). A solution that I can think of is having a 100 ml hourglass shaped flask which is very narrow at the middle. It would be made of quartz, and satisfy the resonance conditions. I think that most of the UV would be able to get through in the very thin middle where the bubble is centered. Only estimations of the temperature have been made by calculating the density of the bubble at its collapsed stage, or from the general visible color.
With my Photo-Multiplier tube, I detected that the flash of light spikes up on
either side of the bottom of the sign-wave on the oscilloscope. The phase of
the flash compared to the drive voltage depends on the power applied, the
water temperature, and the position of the bubble inside the flask. One flash
of light occurs for each period of the sound wave. The width of the light pulse
(according to the articles) is EXTREMELY SHORT, and is much faster than my
detector can respond to, without integrating. The width I measured, is less
than 1/100 thousandths of a second, but that is limited by my detector.
Click here for Oscilloscope Trace Picture
Figure 3: Scope picture with sine wave of the applied voltage, and the light
spike seen by the photomultiplier tube. The x-axis has 10 microseconds/division
time scale, and the y-axis has 0.2 volts/division for the PMT signal.
For the gain I used on my Photo-Multiplier, one photon gives a 10 mV signal. The detector could see about 1 volt from the SL experiment, or about 100 photons. I estimate the number of photons given off by each SL burst: The photo-multiplier tube has an efficiency of about 0.2, my lenses collect only a fraction of the light emitted (1/64), and there are some losses at each glass surface (30% total). Therefore the 100 detected photons, correspond to about 50,000 photons emitted from the bubble, in each burst.
Using pictures I've taken, I figured out that the maximum width of the SL
bubble is only 250 microns wide. I did this by setting up a ratio between the
enlarged picture on my poster and the actual diameters of the flask. Then I
measured the bubble on the picture and proportionalized it into real life. My
accuracy from the photo is about 25 microns. The upper limit of the volume is
6.42x106 cubic microns. Of course, it is actually much smaller, but
this is my crude average estimate. By naked eye, you can barely see the actual
bubble, even when shining a bright light in from the side.
References
1. Scientific American, February, 1995, p. 46-51, "Sonoluminescence: Sound into Light", by Seth Putterman, p. 96-98, The Amateur Scientist, by Robert Hillary and Bradley Barber
2. New Scientist, April 29, 1995, "Bubbles Hotter than the Sun", by Lawrence Crum and Kenneth Suslik
3. Science News, April 29, 1995, p. 266-267, "Inferno in a Bubble", by Jocelyn Kaiser
4. Science News, August 19, 1995, p. 127, "Why Fiery Bubbles Live in a Waterworld"
5. Physics Today, September 1994, p. 22-29, "Sonoluminescence", by Lawrence Crum
6. Physical Review Letters, February 28, 1994, Volume 72 # 9, "Sensitivity of Sonoluminescence to Experimental Parameters", by Seth Putterman, Bradley Barber, C.C. Wu, Paul Robert, and Ritva Lofstedte
7. Physical Review Letters, December 28, 1992, Volume 69 # 29, Light Scattering Measurements of "Repetitive Sonic Implosion of a Sonoluminescing Bubble", by Bradley Barber and Seth Putterman
8. Channel Industries Specification Sheet for the Transducers
9. Hamamatsu Corporation's Photo-Multiplier Tube Specifications Sheet
10. Much thanks to Dr. Lawrence Crum for the pictures and tips!!!
11. Thanks to my dad for getting the parts and being a lowly lab assistant!!!
