In an earlier blog I give a brief history of the glass pipe RAFOS float. One of the really cool attributes of this glass-pipe instrument is that it can be made to follow isopycnal surfaces in the ocean. Up until then all floats, whether the original Swallow float, the SOFAR float, or even the basic RAFOS float, will drift with the water at essentially at the same depth or pressure. But for those of us who worked with floats, designing a float to follow an isopycnal surface as it moves up and down was always high on our list of objectives, we just couldn’t figure out to do this! Yes, we could instrument the float to measure temperature and salinity and adjust the float’s volume to stay at a desired density, but that would be a very costly approach at best. I don’t recall when we realized this, but the properties of the glass pipe allowed us to convert the float from a constant pressure to a constant density (or isopycnal) device.
To see this, consider first the basic Swallow float. If it is pushed away from its equilibrium, perhaps an octopus messed around with it taking it a bit deeper. The float will compress due to the increasing pressure but not as much as the water around it giving it a small amount of buoyancy. When the octopus tires and lets go, that buoyancy will return the float to its equilibrium pressure. That is the basis for all neutrally buoyant float technology. But water parcels vary in depth in time and in space so to follow a fluid parcel as it moves in the vertical, we’d like the float to do likewise. Making a float remain on an isopycnal surface became quite challenge. It’s just as well I can’t remember the details but we started building some kind of Rube Goldberg device of a float, but it went nowhere fast. Very frustrating, I couldn’t figure out how to do this until a shoaling SOFAR float in the Gulf Stream gave us the clue.
The original Swallow float and the SOFAR floats were made of aluminum. It has a smaller coefficient of thermal expansion than that of water. So, if a float drifts into colder water the displaced water will contract more and become denser than the float, i.e., the float becomes lighter and will shoal slightly. We first noticed this when SOFAR floats shoaled as they slipped out of the warm Gulf Stream into the cold slope waters. This gave us the clue: if we don’t want the float to feel temperature, use glass, it has a very small coefficient of thermal expansion.
The borosilicate glass pipe used by the RAFOS float is used for a variety of plumbing applications throughout industry; fortunately, it is also very inexpensive. It has an extremely small coefficient of expansion, much less than that of water. That is exactly what we want: it doesn’t feel temperature and of course glass is completely impervious to variations in salinity. Thus, once its weight is trimmed so it floats on a certain isopycnal (surface), it will stay at that same isopycnal. But the ocean is stratified, so if our octopus pushes the float deeper where the water is colder, it gains buoyancy relative to the surrounding water. When the octopus tires and gives up, the float will return to its equilibrium density surface.
But for this to work we need to match the float’s compressibility to that of sea water. Since glass is very stiff, we add a compressible unit we call a compressee. It consists of a stiff spring inside a watertight housing at the end of which sits a piston that ambient pressure wants to push in, a movement that the spring will resist. The piston diameter and spring constant are chosen such that its compressibility together with that of the float match that of sea water. This is a very simple easy to make mechanical device. The compressees we have designed and used will operate from the surface to ~1000 m depths in the main thermocline. It also serves as ballast: when the float’s mission is finished, it simply drops the compressee so the float can return to the surface and telemeter the data it has collected.