Refraction of Sound

Balloons filled with helium, CO2, or SF6 act as diverging and converging lenses, respectively.

What it shows:

A balloon, filled with a gas different from air, will refract sound waves. A gas denser than air turns the balloon into a converging lens and a lighter gas makes it a diverging lens. An air-filled balloon has little effect.

How it works:

The refraction phenomenon occurs whenever waves travel from one medium to another in which the velocity of the wave changes. The amount of refraction at the media interface obeys Snell's law. Thus, a spherical balloon filled with a gas in which the velocity of sound is markedly different from that in air will act like a spherical lens.

The velocity of sound in sulfur hexafluoride (SF6) is 0.44 times the velocity of sound in air; a balloon filled with SF6 behaves like a converging lens. The velocity of sound in carbon dioxide (CO2) is 0.78 times the velocity in air. A CO2-filled balloon also acts like a converging lens, but not as strong as the SF6. The velocity in He is 2.7 times the velocity in air; a He-filled balloon behaves like a diverging lens.

sound refraction

Setting it up:

The source of sound is a small speaker, emitting a pure tone at 2.5 kHz. The sound is picked up by a microphone (placed 50 to 60 cm away from the speaker) and the signal is displayed on an oscilloscope. When a balloon is held in front of the microphone, one observes its effect by the change in the displayed signal.

A function generator and amplifier provide the signal for the speaker. The PASCO PI 9587C serves this purpose well. The wavelength of the sound should be less than the dimension of the balloon (which is about 12 inches in diameter); 2.5 to 3 kHz works very well. The output of the microphone needs to be amplified before passing it on to the oscilloscope. For regular microphones, the Sure M267 mixer is sufficient. If using the omnidirectional (pressure) Earthworks M30 microphone (shown in photograph), the Yamaha MG10/2 mixer will provide the 48 V phantom power for it. Also notice in the photograph that the speaker and microphone are both mounted high, to minimize reflections off of the cart—this is more important with an omnidirectional microphone.

Having set the volume of the sound to an acceptable level, adjust the gain of the mixer and scope to display a signal which is roughly 1 division peak-to-peak. Hold an air-filled balloon in front of the microphone; no noticeable effect on the signal strength will be observed. Now place the SF6 balloon in front of the mike as shown in the photo; the SF6 balloon will focus the sound onto the mike and produce a full-scale signal on the screen (an increase in signal strength by a factor of 5). Note that in this set-up the focal point is close to the surface of the balloon. Repeat the experiment with a He balloon; it will significantly diminish the intensity of the sound reaching the microphone.


Students often ask whether sound can be refracted. This is a very clear demonstration of the phenomenon. For your reference, the focal length of a spherical lens (measured from the surface of the lens to the focal point) is given by

focal length sphere

where r is the radius of the sphere, n2 is the index of refraction of the lens material, and n1 is index of refraction of the surrounding medium.

If the source of sound is far away, the focal length of a 15-cm radius SF6 balloon is close to its surface. The focal length of a similar CO2 balloon is 19 cm from its surface, and that of a He balloon is -19 cm. For an air filled balloon, the focal point is at infinity (as it should be). Kenn Lonnquist (Colorado State) suggests using difluoroethane (CH3CHF2 in "Dust-Off"), for which the velocity of sound is 0.59 that of air, and thus should work "better" than CO2.