11-10-14: BifwynneThere are two potential effects that brings to mind. The first is skin effect, which results in high frequency currents tending to disproportionately utilize the outer part of a wire's cross section, and underutilize the central part. The net effect being that the resistance of the wire increases somewhat as frequency rises above a certain point. The larger (numerically lower) the gauge of the wire, the lower the frequency at which that effect begins to take place. Although it seems to commonly be overlooked that utilizing a smaller fraction of a larger cross-sectional area will result in the effect being mitigated to a significant degree.
Al, I seem to recall reading that cables made from different metals and alloys conduct electrical signals differently because of the surface (aka skin) characteristics of the particular metal being used. Are you familiar with these concepts?
As can be seen under "examples" in the Wikipedia writeup, "skin depth" is proportional to the square root of the resistivity of the metal being used. So among the three metals that have been mentioned, gold would be the least susceptible to skin effect, and silver the most susceptible. But silver and copper are very close, and gold is not greatly different in this respect either.
Skin effect can certainly be expected to be a significant design consideration in the case of digital cables, as digital audio signals have significant spectral content well into the RF region.
As I see it, though (despite claims that may be made in some cable marketing literature) it will be of no significance in the case of analog interconnects. Partly because for the relatively narrow gauges that are typically used for analog interconnects skin effect will not begin to occur until frequencies that are well above the range of audibility. And partly because a slight or even not so slight increase in resistance won't matter in the case of an analog interconnect, as I indicated earlier.
With respect to speaker cables I recall seeing some calculations in a paper by Bill Whitlock of Jensen Transformers, a renowned expert, indicating that under typical circumstances skin effect may result in a high frequency rolloff amounting to somewhat less than 1 db at 20 kHz (I don't recall what the exact fraction was). I suppose that may be marginally/barely perceptible to some listeners in some systems, but I'd expect it to be overshadowed by other cable-related effects, including (I suspect) those related to the dielectric.
The second possible effect your question brings to mind relates to oxidation and impurities at the surface of the conductor, which as you indicated will vary depending on the particular metal. Those impurities perhaps resulting in diode effects and rectification affecting the small fraction of the signal current which flows through the outermost part of the wire's cross-sectional area. I've never seen any calculations or measurements quantifying that kind of possibility, that would provide a feel for whether or not it has a reasonable chance of being audibly significant. But fwiw my guess would be that since the depth that would be involved is probably a very tiny fraction of the wire's cross sectional area, perhaps even measured in molecules, that kind of effect is likely to be either minimal or negligible compared to other factors in the design of the cable.
Best regards,
-- Al