Anyone know the penetrance of handheld XRF? That will at least tell you the surface composition.
As I said in the last thread about this, I think that a “ringing” test might be the best - tap a gold bar of the same shape as the suspect one with a mallet, and see how it sounds, and then compare.
Don’t know if would work in this case but metal proximity sensing (metal detectors) can be tuned somewhat to particular metals. For gold detection they are tuned not to pick up iron. Possibly they could detect the presence of tungsten based on it’s characteristics in producing eddy currents. A large solid portion of tungsten may be readily detected under plating in that manner.
There are many, many tests you could perform on a sample. For any single test, there’s some way to fool it using the appropriate combination of other materials. The more tests you’re trying to fool at once, the harder it gets.
Offhand, I’d think that it’d be tough to come up with anything that could fool both a test based on density and one based on conductivity. Tungsten is about the only cost-effective way to mimic gold: Most other things with a density in that vicinity or higher cost even more than gold does. But tungsten has a far lower conductivity than gold. In fact, at room temperature, I’m pretty sure that copper and silver are the only materials that beat gold for conductivity, and that by not very much (such that to get the same conductance out of either you’d need to take up most of the object’s volume with them, causing it to fail a density test).
If you wanted to be completely sure, you could also toss in a few different sorts of X-ray tests (crystallography diffraction, bulk transparency, etc.), an MRI test, and so on, but I don’t think it’s really necessary.
Chronos, as I have noted, Isaac Asimov said that in an index of electrical conductivity silver got 100, copper 95, and gold 67. He did not mention tungsten or its related metals. A gauge that shows conductivity would likely identify an amount of metal as being gold or not.
Metal conductivity table here
Wouldn’t the resistance of a bar of solid gold, with a cross section of some cm^2, be so low that it would hardly make a measurable difference if it contained some pellets of another material?
Some form of Eddy current measure should be able to help. The frauds seem to be with ingots that are about 1/16 of an inch gold over a brick of tungsten. The gold is thick enough to pass scrape and ringing tests but the ingot is almost all tungsten. I suspect that adulterating the gold with smaller amounts of tungsten is not worth it. The risk reward not paying off.
Is there any way to make the plating blister? (i.e. a method that wouldn’t damage a solid gold piece).
Does gold plating adhere very well to tungsten? Intuitively, it doesn’t seem to me as though it should (very different hardnesses and I think, not mutually soluble).
Its not usually just plated.
The most recent “case” that probably sparked this concern was a year or two ago. A case, supposedly, in Europe where a solid gold brick sized bar had been drilled out(say half the gold removed) and tungsten placed inside. Then nicely covered with gold where the drill holes in the end of the bar were made. So the weight was still the same, and the density was the same. But the gold shell was thick enough to pass various tests.
Which means that you can get the same conductivity from a mixture of copper and some lower-conductivity material. A conductivity test by itself can thus be fairly easily fooled. The problem is that the materials that could pass a conductivity test wouldn’t pass a density test.
I was also thinking of Ignotus’s point: If you’ve got a 4cm x 6cm bar, with the outer 5mm gold, that’s still (if my quick math is right) 9 cm^2 cross-section of gold. That’s going to have very very low resistance, in fact certainly less resistance than the connection between the bar and any kind of electrical leads you might attach. So I’m not sure how easy it would be to tell the difference between the very very low resistance of a gold-covered tungsten bar, and the also very very low resistance of a pure gold bar, at least using simple conductivity tests.
You don’t use a simple multimeter. The standard technique is a 4 wire meter. You can tune the system to be very good in the range where you need to know the resitance.
Also, a very neat trick would be to simply use a known real ingot as a reference in a bridge.
A fake ingot is likey to have about double the resistivity of a real ingot. As I noted above, it would not be hard to build a device that uses eddy currents to measure the resistivity of the ingot as well. Useful in tha it is non-contact and could screen ingots reasonably easilly.
But would a 4-wire meter be precise enough to distinguish the resistance of 9cm^2 of gold (at maybe 20-cm length?) from the resistance of 24 cm^2 of gold? I’ll take the word of a real EE that it is possible, but it doesn’t seem trivial to me to do with simple test equipment.
Ooh! Ooh!
I’ve got another way!
Gold’s Thermal conductivity is nearly 2x that of Tungsten! Just heat one side of a bar and see how fast the other side gets hot. Should be easy to distinguish Gold from mostly Tungsten.
Hmm, that could work.
What did I tell you? (Of course it was with electricity rather than heat.) 
A device that uses eddy currents to measure the conductivity of an object is exactly what a metal detector is. And in fact, there are a great many metal detectors that are already designed to distinguish between various metals: That’s one of the primary methods by which vending machines distinguish between real coins and valueless slugs.
Are slugs made of steel?
Heap leach it with cyanide …