[Phys-l] magnetism science and engineering

[John Denker <jsd@av8n.com>]

On 08/25/2011 10:33 PM, Bernard Cleyet wrote:
> I assume this is related to the coercivity, JD's explanation is too
> deep for me.
>
> bc has only shallow physics understanding.

That sets BC apart from the folks who have no clue about magnetism 
but think they are experts.

I've been trying to find some online references on the subject.
It's frustrating.  Most of the articles I've found seem to have
been written by people who never even met someone with any real
expertise or experience in the field.

I'm far from being an expert, but I can identify a few topics 
of interest:

1) At the macroscopic level, there are a couple of centuries' worth
 of engineering lore.  Some buzzwords here are "magnetic circuit"
 and "reluctance".  I haven't found any good articles, but you can
 look for yourself:
   http://www.google.com/search?q="magnetic+circuit"+reluctance

 You can use circuit/reluctance ideas in ideal situations and/or
 if all you want are ballpark estimates, but most real-world 
 situations are highly non-ideal so people use finite-element
 methods aka finite-element analysis:
   http://www.google.com/search?q="magnetic+circuit"+FEA+OR+FEM

2) At the moderately microscopic level, the objects of interest
 are the magnetic domains.
   http://www.google.com/search?q="magnetic+domain"&tbm=isch

 At room temperature, if a piece of iron is "unmagnetized" at the
 macroscopic level, it consists of many domains, each of which is
 strongly magnetized, but the domains are not aligned.  When you
 "magnetize" the piece of iron i.e. give it a permanent macroscopic 
 magnetization, you are mostly growing some domains at the expense 
 of others.
   http://hyperphysics.phy-astr.gsu.edu/hbase/solids/imgsol/domains2.gif

 The difference between magnetically "hard" steel (Alnico) and
 magnetically "soft" steel (silicon steel for transformers) has to
 do with pinning of the domain boundaries.

3) At the ultra-microscopic level, we can understand the physics
 of the magnetic field and the atomic physics of magnetic materials.

3a) The physics of the plain old magnetic field (including the
 Poynting energy) is important here.  This by itself explains
 why magnetizing a long thin needle is very different from 
 magnetizing a short fat cube.  If you take a very long needle
 and bend it around to close on itself, you get a "magnetic core"
 which is a particularly favorable geometry with many practical
 applications.

 At the third-grade physics level you can /feel/ that arranging
 bar magnets end-to-end (aligned the same) is energetically favored, 
 while arranging them side-by-side (aligned the same) is disfavored.
 Again this tells you that a long thin needle (bars end-to-end)
 is very different from a short fat cube (bars side-by-side).
 The domains in the cube have an incentive to flip.  A large 2D 
 sheet of bar magnets all aligned the same way perpendicular to 
 the sheet is energetically impossible;  the energy diverges as
 the sheet gets bigger.  More to the point, the energy /per bar/
 diverges.  Hint: scaling argument.

3b) Most of the articles I've found on the atomic physics of
 magnetic materials are frustrating.  Some of them are so sketchy
 that they remind me of pre-schematic "refrigerator art"
    http://highglossonline.com/main/wp-content/uploads/2009/01/augie_art1.jpg

 It doesn't help to explain magnetic materials in terms of the 
 "exchange force" and then they explain the exchange force by
 attributing it to "quantum mechanics".  That explains NOTHING, 
 in the sense that you could use the same words to explain anything 
 and everything.  Everything is governed by the laws of quantum 
 mechanics, so mentioning the magic words "quantum mechanics" does 
 not even begin to explain why 
  -- some materials are ferromagnetic
  -- some materials are ferrimagnetic
  -- some materials are antiferromagnetic
  -- some materials are superconducting
  -- some materials are none of the above

 Also:  I know about gravitational force, electroweak force,
 and strong nuclear force ... but to my way of thinking, there
 is no comparable "exchange force".  Similarly there is no such 
 thing as "degeneracy pressure".

 If you insist on a definition, I would say:
   The "exchange energy" is the amount of energy that you 
   would get wrong if you unwisely analyzed the material in 
   classical terms, neglecting the fact that at the relevant 
   temperatures the electrons are highly degenerate.  

 That definition is intended to be ridiculous.  For a more
 constructive discussion of degeneracy, see
   http://www.av8n.com/physics/degeneracy.htm

3c) This is related to some other rather fundamental physics
 that is widely misunderstood.

 For one thing, consider the notion of filled "Lewis octets" in
 molecules.  This is completely crazy.  Everything we know about 
 the theory (quantum mechanics, atomic physics, molecular orbital 
 theory, etc.) tells us there are no filled octets in molecules. 
 Everything we know about the experimental data (spectroscopy, 
 magnetism, etc.) tells us the same thing.  Good experimental
 data predates the Lewis theory by many decades.  And yet ...
 practically every introductory chemistry book touts Lewis dot
 diagrams as the fundamental explanation of molecular structure.
 Why does anybody put up with this?  This is one of the many
 things that teaches students that critical thinking is not
 tolerated in school.

 There are simpler and better ways of modeling molecules:
   http://www.av8n.com/physics/draw-molecules.htm

 The canary in the coal mine is magnetism, namely the conspicuous 
 paramagnetism of the O2 molecule.  This can be understood in terms
 of Hund's rule, which can in turn be understood in terms of quantum
 mechanics.  If you want to understand what /exchange/ means, you 
 may want to start with this relatively simple system.  It's kinda 
 scary how many people don't really understand this.  Here are some
 notes on the subject, with some possibly-useful diagrams:
   http://www.av8n.com/physics/triplet-singlet.htm

4) The physics of the Curie point is very interesting.  In the
 absence of an externally-applied field it is a second-order
 phase transition i.e. critical phase transition.
   http://www.nobelprize.org/nobel_prizes/physics/laureates/1982/

 Note that if you expand things in a Taylor series centered on the
 critical point you are guaranteed to get the wrong answer ...
 which is an interesting lesson unto itself.

 I haven't found a good tutorial on this topic, either.  You can
 look for yourself:
   http://www.google.com/search?q="curie+point"+"second-order+phase+transition"+"critical+exponent";

=============

Using ideas (1), (2), and (3a) together you can get a passable
understanding of what a magnet keeper does.  For a practical
magnet, the desired configuration of domains is one that produces 
a large magnetic field at the poles of the magnet.  Without the 
keeper this configuration is energetically disfavored.  With the 
keeper it is more favored and hence the domain-configuration is 
more stable.

You can /feel/ the force on the keeper, so you know the "kept" 
state is a low-energy state.  This reduces the incentive for the 
domains to re-arrange themselves into a configuration with less 
macroscopic magnetization.

=========================================

If anybody knows of any good tutorials in this area, please let
us know!