The design and the operation of a continuous flow reflux (CFR) oven used for the formation of an effusive Cs molecular beam which is compatible with ultrahigh vacuum (10−9 Torr) conditions are described. The CFR oven operation utilizes capillary action to recycle the unused part of the Cs molecular beam. A possibility of extending the principle to supersonic beam pressures is discussed.

Electrical transport and optical data on amorphous-metal --- rare-gas mixtures of Cs-Xe at 6 K are presented. The Cs-Xe mixtures fit a model whereby most of the Cs is distributed randomly on the microscopic scale while a small amount exists in the form of granular clusters. Although preferential clustering occurs, it is not prominent enough to apply the model of classical percolation theory to the metal-nonmetal transition which is exhibited. The transition is located at 0.55 \ifmmode\pm\else\textpm\fi{} 0.01 Cs atomic fraction based upon the closure of the optical gap and the threshold of extended-state conduction. The composition range over which the states at the Fermi energy are localized is very small or zero.

Electrical transport and optical measurements are reported on amorphous mixtures of Cs and Au at 6 K. The two metal-nonmetal transitions exhibited on either side of the stoichiometric compound CsAu are studied as a function of composition. The role of the ionic bond in the transition is discussed in relation to the character of the composition dependence of the physical properties investigated. Both the electrical transport and the optical data indicate that the transitions occur at the Cs atomic fractions 0.25 \ifmmode\pm\else\textpm\fi{} 0.02 and 0.60 \ifmmode\pm\else\textpm\fi{} 0.01 and that at compositions between these the states are localized at the Fermi energy. Comparisons are made with data on liquid Cs-Au mixtures.

Optical and electrical transport properties were investigated in amorphous Cs-Sb mixtures at 6 K. A metal-nonmetal transition is observed as a function of Cs atomic fraction ($X$) and both the transport and optical properties indicate that this occurs at $X=0.68$. The dc conductivity at this composition is \ensuremath{\sim} 150 ${\mathrm{\ensuremath{\Omega}}}^{\ensuremath{-}1}$ ${\mathrm{cm}}^{\ensuremath{-}1}$ which is consistent with acceptable values of Mott's minimum metallic conductivity. The mixtures can be partitioned into three distinct composition regions: (1) $X<0.64$, in which the mixtures are semiconducting; (2) $0.640.68$, in which the mixtures are metallic. The optical data indicate that within the broad composition range $0.7

Electrodeless Hall mobility measurements at the metal-non-metal (M-NM) transition in metal-ammonia (M-NH3) solutions are presented. The possibility of electrode effects has been eliminated. These data corroborate previous conventional four-probe data which showed an excess mobility in Na-NH3 solutions near the composition corresponding to phase separation. The possible role of fluctuations on the M-NM transition is discussed in connection with the competing theories of Mott (1974) and Jortner and Cohen (1975). Data from a Li-Na-NH3 alloy indicates a non-linear decrease in Hall mobility with decreasing Na/Li ratio.

An electrodeless (double induction) technique has been developed for the determination of Hall mobility. Measurements were performed on seven solid or liquid metallic samples and the results compared to reported mobilities. With a current of 1 A and magnetic field of 100 G, signals of 100 nV were measured for Cu, Ag, or Au samples, while signals as low as 50 pV can be measured. Hall angles as small as 2×10−6 rad have been determined. The usefulness of this technique in Hall mobility measurements on systems in which electrode dependent results occur or in corrosive materials on which no electrode is stable is emphasized.

The conductivity and Hall coefficient have been measured for solutions of Li, Na, K, and Cs in liquid NH3 using a double ac method over a concentration range of 1–10 mol% metal (mpm). The conductivities vary from 10−1 to near 104 Ω−1 cm−1 and the Hall coefficients from 10−6 to 10−9 m3C−1 over the 1–10 mpm range. Hall mobilities vary from 10−2 to 100 cm2V−1 s−1 over the same range. The mobility for Na–NH3 solutions shows a maximum at 4 mpm while the others are monotonie. The mobility peak in Na–NH3 solutions appears to be related to the presence of substantial composition fluctuations in that system.

We have made a computer simulation of the Hall effect in a inhomogeneous 7 \ifmmode\times\else\texttimes\fi{} 7 \ifmmode\times\else\texttimes\fi{} 21 simple cubic array. Our results are compared with data on Li-N${\mathrm{H}}_{3}$ solutions and effective-medium theory.