In pool boiling, vapor is generated at a superheated wall that is small compared to the dimensions of the pool of nominally stagnant liquid in which it is immersed. The motion of the liquid is induced by the boiling process itself (analogous to single-phase natural convection at a heated wall in an unbounded fluid) and the velocities are assumed to be low. These conditions are convenient for small-scale laboratory experiments and much of the understanding of boiling, such as the basic division into nucleate, transition and film boiling and studies of bubble nucleation and motion discussed in the article on Boiling, has been derived from pool boiling experiments. However, pool boiling is unusual in industrial equipment. Even if there is no forced flow of liquid past the heated wall, confinement of the liquid and close spacing of multiple heaters, as in kettle reboilers, Figure 1, means that conditions are closer to Forced Convective Boiling. The heat source is often a hot fluid separated from the boiling liquid by a thin metal wall, whereas electrical resistance heating is often used in pool boiling experiments. Consequently it is important to appreciate the special conditions of pool boiling experiments and to exercise caution in transferring the information they provide to large-scale industrial systems in which flow effects are generally significant. This article reviews the techniques that are used in pool boiling experiments. Pool boiling behavior is described in more detail in the articles on Boiling, Nucleate Boiling, and Burnout (Pool Boiling).
Pool boiling can be classified according to conditions in the pool, the geometry of the heated wall and the method of heating. These conditions influence the methods used to measure the primary variables of wall superheat and heat flux that are conventionally used to present boiling heat transfer performance as a "boiling curve".
In saturated pool boiling, or bulk boiling, Figure 2(a), the pool is maintained at or slightly above the saturation temperature by interaction with the vapor bubbles rising from the superheated boiling surface. (Subsidiary heaters may be used to compensate for heat lost from the walls of the containing vessel.) The pool has a free surface at which the bubbles burst; the vapor space is usually connected to a condenser that returns liquid to the pool. The system pressure is controlled by the cooling applied to the condenser. In subcooled boiling, the pool temperature distant from the boiling surface is below the saturation temperature. There can be no escape of vapor from a subcooled pool, unless it is very shallow, so a heat sink must be provided by cooling regions on the walls of the vessel, Figure 2(b). Alternatively, a subcooled experiment can be run for a short period without heat sink, relying on the thermal capacity of the cold pool. A subcooled pool cannot have a free surface in contact with its own pure vapor. Either the boiling vessel must be connected to a separate vessel in which the pressure is controlled, or there must be a gas space above the pool. Use of a cover gas leads to a concentration of dissolved gas which can influence boiling, particularly by improving the stability of nucleation sites and reducing the superheat required for their activation. Dissolved gas can be removed by a preliminary period of saturated boiling, either in the experimental vessel or in a separate vessel from which the experimental vessel is filled. The temperature-time-dissolved gas history can influence the subsequent boiling experiments, as described in the article on Nucleate Boiling, and may be different in industrial systems.
In both saturated and subcooled pool boiling, the operation of the heat sink requires a recirculatory flow in the pool that may interact with the boiling process in ways that depend on the geometry of the pool and of the superheated boiling surface. The shape of the vessel may be constrained by the need to observe the boiling process. Early experiments on pool boiling used heating surfaces that were thin horizontal wires of materials such as platinum, heated by the passage of direct electrical current. The electrical resistance of the wire provided a measure of its temperature, averaged over its length. Such experiments are useful to demonstrate some of the basic characteristics of boiling but they suffer from the disadvantage that the length scale of the bubbles is similar to that of the heater so that their behavior is atypical of the extensive surfaces in industrial plant. Most experiments now use larger heaters in the form of horizontal cylinders with diameters in the range 10 to 20 mm, horizontal plates of circular or rectangular shape and vertical or sloping rectangular plates, with dimensions in the range 5 to 100 mm. Heaters much larger than this are rarely used because of the large power requirements resulting from the high heat fluxes in nucleate boiling. The small heaters interact with the recirculation of liquid in the pool through edge effects or because their dimensions are comparable with the critical wavelengths of interfacial instabilities in film boiling, Figure 3. The recirculatory flows that must return liquid right to the wall in nucleate boiling are rarely considered, except in the special case of vertical flow counter to the vapor flow.
Industrial plants frequently use heat transfer from hot single-phase or condensing fluids to drive boiling under conditions that approximate to controlled wall temperature, Figure 4. This is rarely done in pool boiling experiments because of the difficulty of measuring the wall temperature and the heat flux accurately. Electrical resistance heating is generally used, so that experiments are performed at controlled heat flux. This can influence the boiling process, particularly in the departure from nucleate boiling and the transition region, Figure 5. Electrical heating is sometimes combined with sophisticated feedback control in order to operate in the unstable region of the boiling curve where dq/dΔTsat is negative. Electrical resistance heating is used in three ways:
Heaters in the form of cylindrical tubes or rectangular plates can be made of thin, electrically-conducting material. The uniform heat input is calculated from the current and the voltage across the heater. Temperature-measuring devices (thermocouples, thin resistance thermometers or thermochromic liquid crystal) are attached to the nonboiling surface of the wall, which is maintained adiabatic by thermal insulation or guard heating. The measurements are corrected for the parabolic temperature distribution across the wall to give the estimated temperature at the boiling surface, Figure 6(a).
A variation on (a) is to deposit a very thin layer of electrically conducting material such as gold or tin oxide on an insulating substrate such as glass. The electrical heat input is generated at the boiling surface; the temperature is deduced from its electrical resistance or measured at the back of the substrate. When used with a heating layer that is so thin that it is transparent, this technique offers the special advantage that the boiling process can be observed through the wall; the disadvantage is that the wall effectively has the thermal properties of the substrate, which, because it must be an electrical insulator, has a very low thermal conductivity (except for special materials such as sapphire).
Indirect electrical heating is used in conjunction with a thick wall of a good thermal conductor such as copper or aluminum. An electrically insulated resistance element is embedded in, or clamped to the back of, the wall. Thermocouples are embedded in the wall to measure the temperature gradient, from which are calculated the heat flux and the extrapolated temperature at the boiling surface, Figure 6(b). This method can be used for heaters of circular or rectangular crosssection.
Electrical heating is difficult to arrange for other shapes of heaters, such as spheres. For these cases, a transient quenching method can be used in which the heater is treated as a calorimeter. Its temperature and rate of decrease of temperature are measured by embedded thermocouples when the preheated body is plunged quickly into the liquid pool, then held stationary, Figure 6(c). The readings are corrected for the transient temperature gradients within the heater. This method is particularly convenient for the study of transition boiling.
The choice of heating method places constraints on the material and thickness of the heated wall, which may not match the conditions in industrial systems. The role of the bulk properties of the wall is overshadowed by the influence of the condition of its surface. The microgeometry and wettability of the boiling surface are known to have large effects on nucleate and transition boiling, as discussed in the articles on Boiling and Nucleate Boiling. In industrial plant they are dependent on the method of manufacture and on subsequent corrosion and fouling in ways that cannot be reproduced exactly (or even as yet quantified) in pool boiling experiments. Pool boiling surfaces are often subjected to artificial treatment in order to improve the reproducibility of the experiments but, for the reasons indicated above, there is considerable uncertainty in the application of the data to other systems.