Magnetic particle or dye penetrant?

Both find flaws that break the surface. Neither sees anything below it. Beyond that they work on completely different principles, and the difference decides where each can be used.

The rule that settles most cases

MT depends on magnetising the part. If the material will not hold a magnetic field, there is no leakage field at a crack, no particles gather, and nothing appears. This is not a sensitivity limitation to be worked around — the method simply does not function.

A practical consequence: a weld between carbon steel and austenitic stainless can be inspected by MT on one side of the joint and not the other. Dissimilar metal welds usually get PT for that reason.

When both are possible, MT is more sensitive

On ferromagnetic material, properly applied wet fluorescent magnetic particle inspection under UV finds smaller and tighter cracks than penetrant does. Two reasons:

• MT sees slightly subsurface flaws. The leakage field extends beyond the surface, so a crack a fraction of a millimetre below can still be detected. PT sees only what is open at the surface, absolutely.
• MT tolerates a tighter crack. Penetrant has to physically flow into the opening and back out again. Very tight cracks — and SCC is often very tight — can hold the dye so well it never returns to the developer.

MT is also faster. There is no dwell time and no development time; an indication appears as the particles are applied.

Where PT wins, apart from material

• Complex geometry. MT needs the magnetic field roughly perpendicular to the crack, which means magnetising in two directions and thinking about field direction on every surface. Penetrant does not care about orientation at all.
• No power, no equipment. An aerosol kit and a rag. In remote locations or up a structure, that matters.
• No residual magnetism. MT can leave a part magnetised, which attracts swarf and can interfere with subsequent welding. Demagnetisation is an extra step PT never needs.

The mistake that hides cracks from both

Both methods need the flaw to be open at the surface. Anything that closes or covers the opening defeats them, and the most common cause is the preparation itself.

Grinding smears metal across the crack mouth

A ground surface looks clean and inspection-ready. What has actually happened is that ductile steel has been dragged across the opening and closed it. Penetrant cannot get in; particles have no leakage field to gather at.

This is why grinding before a surface inspection has to be followed by etching or another method of removing the smeared layer — and why an inspection performed straight after a grinder is one to be sceptical of. It is a genuinely common way that repair welds are signed off with cracks still in them.

Coating, scale and dirt

Neither method sees through paint. Removing coating is usually the largest single cost in a surface inspection campaign, which is why methods that tolerate coating — ACFM, eddy current array — exist and are worth their premium on large coated structures.

Temperature

Standard penetrant materials have a working temperature range, typically something like 10–50 °C. Outside it, dwell behaviour changes and the inspection is not valid. Hot surfaces need high-temperature consumables or need to cool. MT is less fussy.

What neither of them does

Neither gives you a depth. Both tell you a flaw is there and how long it is at the surface. For a fitness-for-service assessment you need through-wall height, and that comes from ultrasonics — PAUT or TOFD.

Neither finds anything that does not break the surface. Cracking that initiates on the process side of a vessel is invisible to an external surface inspection until it grows through, which is why in-service inspection for sulphide stress cracking or HIC uses ultrasonics from outside rather than MT.

The practical pattern is therefore: a surface method to find it, ultrasonics to size it, and neither one alone is a complete inspection of a critical weld.