Degassing Water

To degass the water I used the “boil the heck out of it” technique in an erlenmeyer flask. To my minor annoyance, the flask I had was just too big for the rubber cork, so innovation involving a rubber band and a bit of plastic bag had to suffice. And suffice it did..
After vigorous boiling for 10 minutes or so, I put the plastic bag over the top and secured it with the rubber band. Boiling like this removes most of the dissolved air in the water. The solubility of air in water decreases with temperature and having a neck that is small relative to the flask means that air doesn’t easily re-dissolve in the water.

This water is cooled and then carefully poured into the spherical flask. Now bubbles should not form all over the place under cavitation – which upsets the acoustic resonance of the flask, now single bubbles can be injected into the water for more fine control.

The first attempt with this resulted in an interesting failure, acoustic resonance was located and all was going well with no ultrasound induced streamers or bubbles appearing. A bubble was introduced to the water by disturbing the surface (design change pending), and stabalised- but not in the center of the flask. A small cloud of bubbles began to form over the surface of the primary bubble, rather like an eclipsing moon, until the primary bubble was made opaque by this tiny foam. This state remained for a few seconds and then suddenly the cloud of tiny bubbles was blown off the primary bubble in a beautiful “smoke” ring. Quite fascinating to watch – almost like the formation of a mini planetary nebula…

The tiny bubble ring then faded and dissolved leaving the primary bubble intact. This was not what I was expecting at all, now looking for some description of this effect. Unhappily it was at this point that the epoxy gave out and the flask detached from the ultrasound transducer. I should like to reproduce the effect and film it for the next experiment.

Trial 2 – Cavitation + Sonoluminescence

In this second trial run I decided to do away entirely with the driver circuit. I think there was a bad impedance mismatch which was limiting the power that could be delivered to the transducer, and being quite happy with the performance of this little amp I decided to try a direct connection to the 100v(max) output of the amp – this made an enormous difference to the delivered power.

The ultrasonic transceiver pair arrived today and I was able to use these like a conventional non-contact microphone in close proximity to the flask (I really need another retort stand..). These are just £2 conventional ultrasonic receiver and transmitters, that you’d find in the old fashioned ultrasonic remote control devices. These are 25khz rated, i.e. their peak sensitivity is around 25khz – which is perfect, you can also get 40khz ones, but wouldn’t be as sensitive, have to locate the data sheet for these..

The amp was wired to the flask with screened cable, necessary since we’re dealing with high frequency, high amplitude signals that will happily radiate into the environment and screw up any measurements.

The flask was filled with ordinary tap water. The output volume of the amp was set to low, and the frequency generator adjusted until acoustic resonance was detected by the “mic” as before. This happened to be around 25khz again, since this is just a trial I’m not recording the specifics. Interestingly, it was found that placing the rubber bung in the flask produced a much stronger acoustic resonance (x3) at the same frequency, than a bungless flask – this flask model resonates very well.

Increasing the amplitude of the signal led to an interesting change in activity in the water, magically strange wispy filament streamers appear around nucleation points in the flask, some appeared in the centre of the flask and danced around. This was accompanied by a terrible screeching wailing sound.


Here is some footage of the effect:

I could see bubbles forming and hovering in the water at certain points, I wondered if there was sonoluminescence occurring at these points. Turning off all the lights and using a little video camera I was able to capture this, probably the most exciting pixels I’ve seen in a while. Actually the footage I took is dotted with these little bright flashes from within the flask – this is multi-bubble sonoluminescence.


The next phase is to attempt single bubble sonoluminescence in trial, then we’ll be able to start doing some proper experiments. To do this I first need to degas some water…

Trial 1 – Plain Tap Water

This is the first trial run, with the flask filled with tap water, untreated in any way.

After tuning the circuit with the water in the flask, we’re achieving electrical resonance at about 26.4khz – this is where the voltage across the transducer is at its peak (the transducer is a capacitor connected across a transformer coil, giving a resonant circuit). To find the resonance frequency I just swept up from 25khz and observed the voltage. I could not do any acoustic flask resonance measurements at this point, having fried my piezo disks. (I have sent off for a 25khz ultrasonic receiver which I think will work better as a mic).

The trace below shows the voltage across the transducer, with the probe x10 attenuated, so each square is 100v, I think we’ve got enough drive voltage with this amplifier now!


And here’s what’s going on in the flask, certainly something.

I noticed little bubbles spontaneously formed and danced around, some seemed to gravitate towards certain points in the flask – these must the antinodes (areas of low pressure). Other bubbles that formed in this vicinity would coalesce into larger bubbles, until buoyancy won out and the bubble would rapidly ascend to the surface. This is all quite promising..

The next step may be to de-gas the water and see if we can get some stable bubbles, I’m pretty happy with the rig, but  I am wondering if it is possible to side step all the degassing and achieve sonoluminescence in plain old tap water. Reading around a bit, Krefting et al (2003) made a study of sonoluminescence in air saturated water, and found that stable bubbles exist in water that has not been degassed. This should be properly explored and documented – certainly if it is repeatable.

  1. Effect of drive amplitude.
  2. Effect of frequency tuning around resonance.
  3. Measurements of bubble position.
  4. SBSL / MBSL

I have videoed one of the dancing bubbles – yes the resolution is all wrong.. but you can see bubble dancing in the field, dispersing, growing and reforming.

Detectors – Piezo Mics

So I’ve realised that these little flimsy piezo disks are just not going to cut it. They practically disintegrate very quickly when driving the flask. I guess they just aren’t meant to be driven by these kind of frequencies… Remember this is just the disk that is epoxied onto the side of the flask to pick up the vibration, this has not been fried electrically – it has been fried acoustically!


A different method of acoustic monitoring could be employed – involving a 25khz ultrasonic transducer that is positioned close to the flask. This way we’ll be picking up the amplitude of the flask indirectly, and not destroying things. The fact that it is not attached to the flask shouldn’t be an issue since the thing is radiating ultrasound which can be picked up some distance away… let’s try it out anyway.

The other thing that turned up was a dead cheap 30W PA amplifier – this cost me a tenner, it’s billed as frequency response up to 18khz, which is some way under what we need, but actually these amps are capable of handling frequencies way outside their spec.


The idea is to drive the amp with 1V signal from the generator and boost the signal for driving the transformer and transducer combination. 30W gives us a lot of headroom – even if we’re talking higher impedances into the transformer, we should get enough power.


The transducer arrived today. This is a 60W 25Khz ultrasonic cleaning transducer. Most of it is metal the interesting piezo parts are sandwiched together in the centre of the device with the terminals hanging off.


I’ve been wondering how to couple the device to the flask, and the answer presents itself quite neatly. The transducer comes with a mounting stud with screw thread so that they can be secured to cleaning tanks and the like via a stud mounting. So I decided to glue the stud onto the flask resonating cavity, and then it would be easy to attach and detach the transducer for any reason by unscrewing it. I am hoping that the coupling between the transducer and flask will be good if they are screwed together firmly – Ed 27/04/2016 – this was a daft idea – the coupling is just not good enough and you get an air-gap vibration – epoxy is the way to go.

Here is a picture of the flask, in case you were wondering what size it was, it is a 150ml flask. I thought I’d mention this as it is not readily apparent from the picture. This was sourced from


The next picture shows the flask with the stud being epoxied onto the flask, and at the same time a piezo transducer which will act as a mic being epoxied to the side of the flask – I think the mic transducer is a little big… we’ll see if this is going to work or not… ( Ed 27/04/2016 – it didn’t work out – see page.)


Next up some experiments..

Driver Rig


What we want is a nice smooth controllable sine wave to allow us to resonate the cavity. The amplitude of the wave needs to be high enough to drive the ultrasonic transducer with sufficient intensity. A couple of ways to achieve this:

a) Circuit resonance – careful design of the driver circuit to achieve maximum electrical resonance of the LC circuit (the transducer having a capacitance) for a given driver voltage. This is nicely shown in the article.


b) Brute force –  using a powerful audio amplifier and signal source and a transformer  to generate the necessary power.

Driver Circuit

It was probably best to decouple the signal using a simple transformer, with this method we do need to protect expensive equipment against electrical surge and current as the transients can be damaging. The components for my driver circuit are:

4 x TVS Diodes
1 x Gas Discharge Tube
1 x Resistor
1 x Transformer 6VA 0-6v 0-240v

At this point I am not sure if the 6VA transformer will be up to the job, I am still waiting for the transducer to arrive so I can’t test it; if it burns out I will upgrade it with something better.


The TVS diodes are transient voltage supressors. They break down when the voltage exceeds a certain amount and conduct, shorting out the spike and will afford some protection to connected instrumentation from transients generated from the transformer coils.

The Gas Discharge Tube does the same thing for high voltage spikes and is probably complete overkill – but I really don’t want to blow up my signal generator and scope, maybe I should have stuck one of these across the output.

Finally the resistor reduces any current surge that might occur – again probably overkill. The images below show the bits and the assembled circuit board, which actually IS a bit of board with the components glued on. The terminals are just little nuts and bolts you can pick up from the £($ US) store.


WARNING: when driving this circuit – the voltage on the primary coil of the transformer is likely to be in 100s of volts = unpleasant shock hazard – don’t touch.
Rules of thumb, don’t touch any bare wires, hook up equipment before switching things on, make sure no clutter around that might get in the way, make sure cats, gerbils, young children are not in the vicinity and cannot interrupt, (cats love jumping up on workbenches at critical times, young children fiddle with everything despite your best attempts to tell them not to, can’t speak for gerbils), just be sensible.

Next we’ll hook up the signal generator directly, get some measurements and see how this little circuit performs on its own into no particular load attached – since the load ( the ultrasonic transducer) is in the mail somewhere.


The circuit was connected to the frequency generator and scope with a x10 attenuation probe, and the amplitude of the driving frequency dialled up to it’s maximum. The frequency was set between 15 and 25Khz.

Here is the output with around 25kHz driving the circuit using a x10 attenuation probe. We have 5.0v x 4 squares on the scope x 10, so we’re achieving a good 200V peak to peak output at this frequency.


Here is the output with around 15kHz – the scale is set to 10.0V squares on the scope with x10 attenuator probe – now we’re achieving around 10.0v x 4 x 10 400V peak to peak.



As we decrease the frequency down from 25khz the voltage goes up, if we increase the frequency from 25khz the output voltage drops down, except around 26khz there is a brief increase in amplitude over a few Hz – this must be the electrical resonance point of the circuit.

There is not much else to be done with this circuit now until the transducer shows up, and we can guage what kind of power it will deliver – the resonance point is going to be different with the transducer attached.

Sonoluminescence Experiments


This is the beginning of the project, designing first what the prototype is going to look like, dusting off some old pieces of equipment, finding bits of wire and accumulating some inexpensive parts from Ebay.

The research question for this first project is simple enough – “can I achieve SBSL?”. This breaks down into the following sub goals:

  1. Can I build a decent stable and controllable rig.
  2. Can I achieve a good resonance of a flask without any nasty ripples.
  3. Can I successfully source and degas water for use in the experiment.
  4. Can I suspend a bubble in an acoustic field.
  5. Can I excite that bubble enough to reach SBSL.
  6. Can I detect that visually and acoustically.

Background Reading

There are a lot of papers out there on the phenomena, there are several good books, and a handful of good websites, I’ll collect all the references in one place with links rather than list everything in the Blog.

Books to Read
Sonoluminescence by Young
The Acoustic Bubble by Leighton
Shock Focussing Effect in Medical Science and Sonoluminescence by Srivastava et al.
Cavitation and Bubble Dynamics by Brennen.

Papers to Read
I’m not going to list all the papers here, suffice it to say there are oodles of them and I’ll probably come back to these once this first experiment is a success.. or failure..

Sites to Visit
Techmind as a nice write up:

Physwiki  has some good background information

UCLA Putterman Research Group for some inspiration – this is how the science professionals do it.

Sound Into Light

And finally an interesting site called Dans Nerdery with an unorthodox and amusing account of his own explorations sometime last year, which by his account were entirely successful. He has some nice ideas about controlling tuning of the system which we’ll get to later.

Rig Design and Construction

  1. Design a simple rig that might work.
  2. Source suppliers and parts
  3. Build the rig, with close attention to safety – 4 year old and cat proofing.

To kick off we’re looking at this set up:

  1. Oscilloscope – replaced the old phillips scope which smoked, with a non smoking DS1052E.
  2. Signal generator – a few to try out – but will probably settle on the MHS5200 as its got fine control.
  3. Amp – of some sort
  4. Driver circuit
  5. Oscillation Vessel
  6. Transducer
  7. Mic
  8. A box of instrumentation things for analysis – we’ll get to this later!


Some pictures below of the kit, I decided not to use the old Farnell generator because the signal wasn’t particularly stable in amplitude, and opted for a modern digital frequency synthesis box that can be gotten off Ebay for a reasonable price. The user controls are a bit odd, but fine once you get used to it, it is controllable from the USB port if you have a mind to tinker using a PC.


And the trusty Rigol DS1052E – does what it says on the tin without complaint – good little scope for this project, brilliant for getting trace pictures off and into this blog, plus potential for USB control and tinkering.