Tungsten aka Wolfram is paramagnetic so it is weakly attracted to magnets.

A guy devised the following to test for Tungsten in gold bars:


But he is using a paperclip and not real Tungsten.

Question: How strong a magnet do I need in order to achieve a significant attraction to Tungsten. By significant I mean that it should exert a force of several (1/100)*grams so that it can be measured with a scale as in the YouTube video. Will a Neodymium magnet do? If yes, what should be the strength of the magnet in kg?


It is true that gold is diamagnetic and tungsten is paramagnetic.

The "magnetic susceptibility" of the materials is needed to quantify the effect.

Here is a table of magnetic susceptibilities:


The problem I see with this technique is, as you can see by looking at the values in the table, that there are substances having susceptibilities hundreds or even thousands of times the absolute value of susceptibilities of gold and tungsten, not to mention ferromagnetic materials. In other words, a small impurity in the gold could give the same result as the entire bar being tungsten. The bar could be 99.9% pure gold and give the same result as 100% tungsten depending upon the other 0.1%.

Also, keep in mind that when tungsten faking has actually occurred, the gold bar has been drilled out and only a percentage of the gold bar has been replaced by tungsten. If, say 30% of the gold is replaced by tungsten, you will not observe a net attractive force on the bar, considering the weighted average of the opposite-sign susceptibilities of gold and tungsten.

There are two other important issues with the technique. Firstly, oxygen is paramagnetic. As you can seen in the table, oxygen has a much greater molar magnetic susceptibility than tungsten. When the gold bar is placed above the balance, air containing paramagnetic oxygen is displaced. This effect must be considered. Second, the magnetic susceptibility of the plastic envelope must be considered.

Also, the gradient of the magnetic field (how the field's intensity varies with space) determines the force, rather than the strength of the magnetic field. Magnetic susceptibility experiments are often part of undergraduate physical chemistry lab course. Typically, the sample is suspended from a microgram balance, the sample being between the poles of a large electromagnetic.

Checking for ultrasound echos coming from the interior of the bar at a gold/tungsten interface and measuring the speed of sound in the bar are techniques in use to rule out gold plated tungsten bars.

  • $\begingroup$ Would testing for electric resistance be a possibility? $\endgroup$
    – Roland
    Mar 11 '14 at 19:18
  • 1
    $\begingroup$ Gold conducts about twice as well as tungsten. The challenge would be how to make the resistance of the bar a significant part of the resistance of a circuit. Resistance is inversely proportional to cross-sectional area, and gold bars have very large cross-section compare to normal wires. The bar would make a trivial contribution to resistance whether gold or tungsten. You could distinguish a long, thin gold wire from a long, thin tungsten wire, but trying this with a bar would be very difficult. $\endgroup$
    – DavePhD
    Mar 11 '14 at 19:40
  • $\begingroup$ The link doesn't work $\endgroup$
    – M.M
    Jan 21 '19 at 9:55

The best method of detection is to measure the actual conductivity of ingots. For this, it is necessary to measure the electrical resistance and the geometric dimensions of the object, and this with extreme precision.

In particular, the measurement when the tungsten content is in powder form, is detected and is not detected with the processes using the ultrasonic reflections.

The process is good enough that it has been commercialized and test systems can be purchased off the shelf.

  • $\begingroup$ Can you name such a system please? Link? $\endgroup$
    – Roland
    Apr 17 '20 at 13:01
  • $\begingroup$ Well, your answer has nothing to do with the question, which was whether it's possible to do it with magnets. $\endgroup$
    – Lodovik
    Feb 17 at 23:07

Not the answer you're looking for? Browse other questions tagged or ask your own question.