Liquid metals experimental considerations

Summary of experimental data Film boiling correlations have been quite successfully developed with ordinary liquids. Since the thermal properties of metal vapors are not markedly different from those of ordinary liquids, it can be expected that the accepted correlations are applicable to liquid metals with a possible change of proportionality constants. In addition, film boiling data for liquid metals generally show considerably higher heat transfer coefficients than is predicted by the available theoretical correlations for hc. Radiant heat contribution obviously contributes to some of the difference (Fig. 2.40). There is a third mode of heat transfer that does not exist with ordinary liquids, namely, heat transport by the combined process of chemical dimerization and mass diffusion (Eq. 2-162). 

There is considerable evidence (D3, G7, PI, P4, SI) that bubbles in liquid metals show the behavior expected from studies in more conventional liquids. Because of the large surface tension forces for liquid metals, Morton numbers tend to be low (typically of order 10 ) and these systems are prone to contamination by surface-active impurities. Figure 8.10a shows a two-dimensional nitrogen bubble in liquid mercury. For experimental convenience, the bubbles studied have generally been rather large, so that there are few data available for spherical or slightly deformed ellipsoidal bubbles in liquid metals. Data... 

In 1958, Pitzer (141), in a remarkable contribution that appears to have been the first theoretical consideration of this phenomenon, likened the liquid-liquid phase separation in metal-ammonia solutions to the vapor-liquid condensation that accompanies the cooling of a nonideal alkali metal vapor in the gas phase. Thus, in sodium-ammonia solutions below 231 K we would have a phase separation into an insulating vapor (corresponding to matrix-bound, localized excess electrons) and a metallic (matrix-bound) liquid metal. This suggestion of a "matrix-bound analog of the critical liquid-vapor separation in pure metals preceeded almost all of the experimental investigations (41, 77, 91,92) into dense, metallic vapors formed by an expansion of the metallic liquid up to supercritical conditions. It was also in advance of the possible fundamental connection between this type of critical phenomenon and the NM-M transition, as pointed out by Mott (125) and Krumhansl (112) in the early 1960s. 

A considerable amount of experimental and theoretical work is available on the heat transfer coefficient for turbulent flow in channels. Through most of fhe correlations were developed based on circular pipe, with appropriate corrective factors, the correlations can be applied to various geometries. The heat transfer coefficients at the entrance region of the channel are higher. For nonmetallic fluids (Pr > 1) the laminar layer is very thin compared to the turbulent region, and the heat transfer coefficient is less sensitive to boundary conditions. For liquid metals (Pr < 0.4), conduction heat transfer is important and the boundary conditions have an impact on the heat transfer coefficient. For fully developed turbulent flow of nonmetallic fluid the heat transfer coefficient is expressed as ... 

Best estimates of the experimental values for the electrical resistivity in the liquid, Pl> solid, and of dpJdT for the R s at their melting points, are shown in table 9. Similar values for the thermoelectric power, Q Q, dQJdT, and (<2l Gs) displayed in table 10. Considerable difficulties have been encountered in containing these highly reactive liquid metals, and in maintaining their purity while their electronic properties are measured. For discussions of the ways these difficulties have been addressed, the reader is referred to the literature references cited. 

While the few examples quoted provide some general guidance as to the behaviour of nickel-rich materials in contact with molten metals and salts, it cannot be over-emphasised that such behaviour can be very considerably modified by the presence of very small amounts of contaminants in the liquid media (see Sections 2.9 and 2.10). The effect of very small contents of sodium chloride on the corrosion of nickel-base alloys by sodium sulphate has been referred to previously and other reported examples involving trace amounts, particularly of gaseous impurities, underline the need for great care in interpretation of experimental results. 

There are few methods which can measure well-defined metal fractions with sufficient sensitivity for direct use with environmental samples (approach B in Fig. 8.2). Nevertheless, this approach is necessary in the experimental determination of the distribution of compounds that are labile with respect to the time scales of the analytical method. Recent literature indicates that high-performance liquid (HPLC) and gas chromatographic (GC) based techniques may have such capabilities (Batley and Low, 1989 Chau and Wong, 1989 van Loon and Barefoot, 1992 Kitazume et al, 1993 Rottmann and Heumann, 1994 Baxter and Freeh, 1995 Szpunar-Lobinska et al, 1995 Ellis and Roberts, 1997 Vogl and Heumann, 1998). The ability to vary both the stationary and mobile phases, in conjunction with suitable detector selection (e.g. ICP-MS), provides considerable discriminatory power. HPLC is the superior method GC has the disadvantage that species normally need to be derivatised to volatile forms prior to analysis. Capillary electrophoresis also shows promise as a metal speciation tool its main advantage is the absence of potential equilibria perturbation, interactions... 

---