SOUE News Issue 4

The 17th Jenkin Lecture, 23 October 2004: Bubbles

Dr David Kenning - report by David Witt

Almost the first point in David's lecture was that there were really two quite different sorts of bubbles. To the layman, the image first suggested by the word may be the spherical object floating in the air above the sink when someone has been having fun and games with the detergent. Such a bubble is of course a quantity of air enclosed within a much larger volume of air by a thin film of liquid. For it to be stable, the liquid has to be a mixture, e.g. of water and soap or detergent. The principal engineering significance of such bubbles is when they congregate in large numbers to make "foams".

The other sort of bubble is a quantity of gas or vapour inside a larger volume of liquid, as seen for example when the top is taken off a lemonade bottle, or in a boiling kettle. It was the particular role of bubbles in the boiling process that had been the subject of David's research for many years, and it was two aspects of this that formed the main topics of his lecture:

  1. better methods for cooling electronic equipment;

  2. the role of micro-bubbles in steam explosions in nuclear reactors.

But first we were given a short exposition of the "physics of bubbles". The balance of forces across the surface of a bubble is such that it is in equilibrium when the surface tension of the liquid balances the pressure difference between the vapour inside and the liquid outside. In "nucleate boiling", the most commonly occurring form, bubbles of vapour form on the heated surface, preferentially at small cavities. Vaporisation occurs mainly on the "triple contact line", where the vapour-liquid interface meets the heated solid surface. If the surface is easily "wetted" by the liquid, the bubbles tend to leave when they are quite small. On a poorly wetted surface, which "prefers to be dry", the bubbles cling on until they are much bigger, and the surface under them may overheat.

Electronic equipment dissipates heat, which has to be got rid of. In modern equipment the heat dissipation per unit area is getting rather large, whether in high-power semiconductors (diodes, gate-turn-off thyristors etc.), or in the central processors of computers as they get faster and faster. It is a question of removing heat continuously from a very intense source, and disposing of it to a diffuse sink, usually the atmosphere. So far, the heat has usually been removed by air convection, either natural or forced with a fan, but this technology is approaching its limits. And for some applications, the power required to drive the fan, or the noise it makes, are major drawbacks. David threw out a target of 2 MW/m2 (or 2 W/mm2)1 for rate of heat removal from a semiconductor at 125°C, and started to investigate whether it might be got with boiling heat transfer.

"Heat pipes" are an existing form of this technology, currently used in lap-top computers. The liquid in the pipe is evaporated at one end and condensed at the other, then returns by capillary action via a wick on the wall of the duct.

But it is possible to do without the wick, which is a rather slow way of moving liquid. The geometry first considered was a rectangular channel of cross-section 1 mm x 2 mm, heated at one end and cooled at the other. Bubbles of steam, separated by "slugs" of water, move in one direction, and water gets back again by flowing past the bubbles in a thin film on the walls. With a heat flux of 2 MW/m2, it is fairly easy to calculate that the depth of the water film in the heated section will be decreased by evaporation at a rate of about 0.8 mm/s, if not topped up in some way. But it was predicted that the film under the bubbles would only be about 7 µm thick at the most anyway, so if not "topped up" would disappear completely in about 9 ms. So things have to happen quite fast. There will be very rapid accelerations and high transient pressure fluctuations.

The process is not yet well understood, and there are many design variables to be selected, so a combination of computer modelling and experimentation is under way in many places, sometimes with channels much smaller than the 1 mm x 2 mm just considered. The results of computer modelling do not yet match experiment as well as could be hoped, so the models are clearly in need of improvement. But the target of 2 MW/m2 looks as if it may be achievable.

David then went on to consider his number two topic: what happens when one very hot liquid suddenly mixes with another cold, volatile, liquid. This can occur under extreme fault conditions in a nuclear reactor (and did at Chernobyl). Less serious incidents can happen in a chip pan, and more serious ones in volcanoes, such as Krakatoa in 18832.

Consider the case of very hot molten metal falling into water. If the metal stays in fairly large drops, say 10 mm or so in diameter, then the heat transfer is by "film boiling", in which an unbroken vapour layer forms around the drop. Heat transfer across this layer is quite slow, and thermal equilibrium will be reached in a reasonably "controlled" manner. But if some disturbance breaks up this film, the boiling rate increases dramatically. The resulting pressure pulse can break up the drop into numerous smaller drops, thus greatly increasing the surface area and boiling rate. If the disturbance then breaks up neighbouring drops in a chain reaction, then there can be a "steam explosion".

In an experiment with a hot sphere in water at 100°C, it was found that film boiling occurred if the sphere was hotter than 300°C, and nucleate boiling (the faster version) if it was cooler than 150°C. In between was transitional. So most of the heat in very hot metal would come out fairly slowly. But if the bulk temperature of the water was lower, e.g. 50-70°C, then a different phenomenon appeared, "micro-bubble boiling". Small bubbles appeared on the surface of the hot sphere, but did not coalesce. Instead they floated away and condensed in the surrounding water, thus giving rapid heat transfer, even from very hot metal.

The lecture concluded with a brief look at some medical and biological applications of bubbles. One of these, "lithotripsy", the fracturing of kidney stones with collapsing bubbles generated by focussed ultrasound, is already in use. Other applications are perhaps around the corner.


David Kenning spent 40 years (1963-2003) on the academic staff of the department, most of them also as engineering tutor at Lincoln. He is continuing active research in his retirement, at Brunel University.

The Jenkin Lecture (and the Society's AGM) were preceded by two half-hour talks given by relatively recent graduates, both of the 1995-9 vintage and both from St Catherine's.

Claire Edwards (née Lewis) went to work at a Corus steelworks in South Wales after graduating, until recently, when she moved to pharmaceuticals at Glaxo-SmithKline in Kent ("from one production line to another" as she put it). In "Developing the Engineer" she talked of her experiences, and of the need to keep learning, and not just about technical matters! She made the point that people at all technical and professional levels were necessary for the successful functioning of any industrial activity, and expressed the view that to create artificial distinctions between different levels was misguided. Engineering Institutions please note! (Universities too perhaps?)

Gemma Long has recently been training as a patent attorney (like a rather surprising number of our graduates over the years), and spoke about "The Law of Invention", with which she is rapidly becoming acquainted. How inventive does an invention have to be? And who is the "skilled man", to whom a new idea may, or may not, be immediately obvious? She followed up this look at the law by entertaining us with an account of some remarkably curious inventions that have been patented, or at least offered for that purpose, over the decades.


<<   Previous article Contents Next article   >>