Sent: Wednesday, February 13, 2002 11:20 AM
Subject: Re: [JN] Re: mumetal


My curiosity about mu-metal and ferromagnetics was recently stimulated by my discovery of some left-over mu-metal in our shop. Almost all of what Stu writes is correct, but perhaps a little misleading and inexact. 
Please indulge me if I get a little pedantic here...

Mu metal is a "soft" ferromagnetic material, as opposed to "hard" ferromagnetics that retain a macroscopic internal field after removal of an external magnetizing field. Most alloys are about 80% Ni, near 20% Fe, some Mo and I presume Co in specific alloys. Low C or "mild" steel
is also a soft ferromagnetic material, and thus is suitable for magnetic shielding. Construction grade mu metal, like that used here in our MEG machine room, has a magnetic shielding effectiveness  100x to 1000x higher than mild steel for equivalent thickness.

Ferromagnetism originates at the quantum level. Ferromagnetic elements have a lowest energy e- orbital state that aligns the e- spins parallel, thus giving an intrinsic magnetic moment to the atom. The lowest microscopic energy state of an ensemble of atoms is, somewhat counterintuitively, with these magnetic moments aligned to produce a net magnetic field. Since maintaining an external magnetic field would require energy, the lowest macroscopic energy state requires that the atoms divide into domains about 1000 atoms across and that the magnetic orientation of these domains be randomized.

In the presence of an externally applied magnetic field, the domains  magnetically realign to some degree and thus generate their own external field. The primary field continues to exist, but now the net field is the sum (or superposition) of the primary and induced field. The induced field must have an alignment of _opposite_polarity_ to the primary field (just like two bar magnets must align N to S and S to N), and the superposition of the two fields results in a lower observed
field - voila, magnetic shielding. 

The position and alignment of the shield matters. Shields aligned parallel to the "lines of force" from the source will be much more effective than when aligned perpendicular. You can even shield by placing the shielding beyond the shielded object, although since the source field falls off with distance, the induced field in the shield will be smaller and less effective.

Regarding annealing, the manufacturers claim an increase in magnetic permeability mu of about 40x by hydrogen annealing. The annealing serves two purposes; grain alignment to lower the reluctance, raising the mu, and the removal of impurities such as C. Working the material
can disturb the grain boundaries and increase reluctance, but I'd guess that this effect only matters in the most critical applications, which must be annealed after fabrication. I'd also guess that even the most roughly handled, whacked out mu metal would still be an orders of
magnitude more effective shield than mild steel.

BTW copper roof valley flashing is cheap, solderable and can be cut with scissors - likely much cheaper than a plane of copper tape.

hth tr
Visit MRI, tube audio, reptiles
Tim Reese, MGH NMR Center

Stu wrote

>The stress induced in Mu metal needs to be annealed by a pure hydrogen bake
>out at about 1200 degrees. Not something you'd want to attempt at home and
>probably is the chief reason that mu metal is so expensive. The metal itself
>is not anything really pricey, nickel mostly. Mumetal needs to have a nice
>radius to all edges or it will impart a harshness in the upper frequencies.
>Bending radius should be minimum 4 times the thickness.
>You can create electrostatic shielding by lining the mu metal with 3M's
>conductive adhesive copper foil tape, although it is not cheap (Digi Key
>carries it).
>Magnetic lines of force would rather travel through ferrous material than
>through air. If shielding is what you want, think about transformer cores.
>Many thin layers work better than one thick layer in diverting the lines of
>force. For DIY use, mu metal foil is recommended and annealing is not
>required. A pair of snips or scissors will work well for cutting the foil.
>Try Magnetic shield corporation ( )or Mu shield (
> )as a source. I believe Michael Percy gets his product
>from one of these companies. Both have kits and gaussmeters available and
>both sell foil and sheet in various gauges and widths. IME, simple iron
>works quite well for shielding purposes.