Pengwuino : An alloy of A,B,C,... will have a thermal conductivity, K(A,B,C,..) < K(A), where K(A) > K(B) > K(C) > ...
In other words, you can only say for certain that the alloy will have a thermal conductivity that it poorer than the thermal conductivity of the best conducting component. But in most cases, the thermal conductivity is lower than that of all components. The exceptions usually happen at very small alloying levels where the mean separation between impurity particles is greater than the electron/phonon mean free path. Eg : 0.1% Cd in Cu has K = 377 W/Km which is greater than that of cadmium)
Astronuc : Where did you get that data for Zircaloy ? Page 12 from the link below gives different numbers.
erickalle : I don't see how your question is related to Wiedemann-Franz. It is talking about heat capacity rather than thermal conductivity.
As for what you are saying, there are a few errors in your understanding. A free electron has only 3 (translational) degrees of freedom (rotation of a point particle takes no energy), and so can gain about (3/2)kT of thermal energy (not 3kT). To say that each
of the free electrons can have this energy is wrong though - and that is a failing of the classical picture. The quantum statistics of the free electrons dictates that only a small fraction (~ 1% at room temperature) of them can actually gain this kind of energy. This follows from the Fermi distribution for particles that obey Pauli's Exclusion Principle.
Additionally, what you talk about is called the electronic heat capacity, and is only a part of the total heat capacity. The rest of it comes from the lattice of positive ions. The electronic heat capacity does scale with the number of electrons per atom.