Double Diffusion Experiment


A "double diffusive" process is one that is driven by the presence of two different substances which diffuse at different rates. The salt fingers instability is one example of such an effect. Another example is the effect that is observed when a mass of water with a vertical salinity gradient is cooled or heated from the side.

Consider a region of the ocean with a continuously-varying vertical salinity gradient (saltier at greater depths, i.e. stable density stratification) and at uniform temperature. What is the effect of introducing a cooling influence from the side by, say, an iceberg? Since heat diffuses so much faster than salt, we may regard the salinity of any given parcel of water as constant. The water nearest the iceberg, upon being cooled, becomes dense and begins to fall. Since the density of the far-field water increases with depth, the cooled water soon reaches a depth at which its density matches that of the warmer, saltier water away from the ice, and so its descent is halted. As parcels of water above this one imitate its motion, it is displaced horizontally away from the ice.

The net effect is that several horizontal layers of motion develop, each with nearly uniform density. The relevance of this effect is that the presence of an iceberg may be felt at surprisingly long distances away from the iceberg itself.

We demonstrate this effect in the following experiment. If the above explanation of the physical process is correct, then the thickness of the layers should depend upon the amount of stratification. We thus measured the thickness of the layers for different stratifications.

Experimental Set-up:

A 50cm by 50cm by 10cm tank was filled with salt water such that there was a vertical density stratification. The concentration of the salt was approximately 10% at the bottom and half that at the top. (This was accomplished via the "double bucket" technique.) A 30cm by 5cm by 3cm "iceberg" was prepared beforehand by freezing a solution of dye and water. The iceberg was inserted into the water along the side of the tank and was fixed into place. As the layers started to form, potassium permanganate crystals were dropped into the water to leave streaks by which we could visualize the layers.


Qualitatively, it was observed that the layers formed almost immediately after placing the ice block into the tank. Also, the layers extended all the way across the tank (see image and/or movie below).

The experiment was performed twice, the second time with a smaller density gradient so as to observe the dependence of the layer thickness on the density gradient. Our measurements did indicate that the smaller density gradient led to thicker layers, which is consistent with the theory explaining the process.

Our experiment sustained a "happy accident". On our second run, we positioned a small lamp near our tank to provide improved lighting. We subsequently noticed several double diffusion layers forming that did not seem to come from the ice block (see the picture below). They were significantly smaller than the other layers, and were positioned below the ice, near the light source. The incandescent bulb acted as a heat source, and these layers were the result of that heat! This clearly demonstrated how sensitive stratified water is to side heating or cooling.
An ice block melting in salt stratified fluid
An mpeg movie of this experiment is available.

Experiment performed and written up by:
Paul F. Choboter, Sept. 98,

Thanks go out to Adarsh Mehta for her help in conducting the experiment.
Page constructed by:
Bruce R. Sutherland, Jul. 98,