Nature already did it, in form of neutron stars or "magnetars". Magnetic fields of order 1010 T (or even higher) are present there. This is possible because of very high density of neutrons, together with sufficiently low temperature. Out of reach in Earth-based labs.
If you just simply join the permanent magnets to increase the magnetic field strength then it will not just add and give you the net magnetic field however you have to follow a definite pattern to join these permanent magnets so that you can add up the magnetic filed lines and get a very high and uniform magnetic field which is called Halbach formation .
But disadvantage of this technique is that by just joining these permanent magnets forcefully will reduce its strength .
Stacking two magnets together simply makes a bigger magnet - yes the overall field is larger and at a given distance you have a greater torque on a magnetic dipole (or a greater force on a moving charge) but the intensity of the field (if anything) is smaller (you've got some non-ferrous plating on the magnets perhaps - and so the overall magnetic moment is smaller).
So the H field is larger - which is trivial to arrange.
The thing you are overlooking is self-demagnetization. This is the adverse magnetic field created inside any magnet with free poles. This is why the magnetic field at the surface of a magnet is just a fraction of the Br, depending on the dimensions of the magnet. The equation governing this effect for an axially magnetized disc can be found in the attachment, when x=0. For a very long rod, the equation tells us that the best magnetic field we can hope for is 0.5 Br. If we bend the magnet around like the letter "O" and cut a tiny slot in the magnet, we can get a field essentially equal to Br, but only in a very tiny space, which is not very useful. If we make the slot bigger, then the field falls off, following Maxwell's equations. So I don't think it is possible to get a field more than Br with just a rod, or a large number of them.
As mentioned above, the Halbach arrangement, can do a bit better than a simple rod in creating a magnetic field. There is some focusing and superposition of the fields from each contributing magnet, which is responsible for the effect. But the field at the center of a Halbach dipole is proportional to the natural log of the radii. So again getting 1000 times more field isn't realistic, since e^(1000) is a very large number. (Neither my calculator, nor Excel were happy with me when I tried to calculate it!)
Hope this explanation helps.
Regards,
Stan Trout
P.S. When I learned about self-demagnetization from Chad Graham many years ago in graduate school, he told me that self-demagnetization is one of the most difficult concepts in magnetism. I've carried that wisdom a long way, and still believe it is true today. So if you can figure out this concept, you'll be way ahead of most people.
mi...magnetic dipole moment i (=Spin at position i)
V...volume
Now rewrite the Sum(mi) to a volume Integral over a dipole-density-function.(Of course here are some troubles because spins are arranged with a certain distance in between. Consequently the concept of a "density" is rather doubtful, but the introduction of a coarse grained density is always possible)
For simplicity, assume a constant dipole-density-function, then the Magnetization becomes independent of the volume and is simply given by,
M=(dm/dV)
(dm/dV)...dipole density
Accordingly, stacking or combining of individual magnets to a big one, does not increase the magnetization.
If discussing about the spontaneous magnetization of a magnet (permanent) you can consider your goal the density of magnetic dipoles per unit volume, just like Markus described previously. The magnetic dipoles are given by the spin of the electrons in the electronic layer of an atom. An atom can get to have many, many electrons in its orbit, but the spins of these electrons get oriented (up, down) such that most of the dipoles are canceled. The only elements to have uncoupled spins to show a workable spontaneous magnetization at room temperature, and be sufficiently abundant in nature are the elements in the 3d category: notably Fe, Co, Ni. Follow this link to see the arrangements of the spins in the orbits: http://www.ptable.com/#Orbital
The experiment you proposed to take many electrons and align them to use their magnetization is what is used today to generate magnetic fields in electromagnets. The thing is that you need to align the spins of these electrons . Easy peasy: use an electric field. The problem: the electrons will travel along the electric field. Continuation of the problem: the variation of an electric charge in time through a cross-sectional area = electric current. The higher the desired magnetic field strength the higher the current. With high current comes increased temperature due to the electrical resistivity of the materials (Cu, Al) employed to 'store' the electrons, actually these materials only allow the passage of the electrons. For very high magnetic field levels you need to go super-conducting. Super-conducting = expensive materials, expensive cooling, poor mechanical properties of the materials, etc., etc.
Nature already did it, in form of neutron stars or "magnetars". Magnetic fields of order 1010 T (or even higher) are present there. This is possible because of very high density of neutrons, together with sufficiently low temperature. Out of reach in Earth-based labs.
James Garry said: "Stacking two magnets together simply makes a bigger magnet"
If I stack two magnetic dipoles atop, then the magnetic flux denisity will be twice as high, so will be the B-field. If aside - the same.
Max said:
>Electrons act as small magnets and many of them combined and their >magnetic fields aligned together produce a powerful magnet.
> Can't this be done macroscopically?
Yes. Just stack two flat magnets atop. You could reach up to 0.5-1 Tesla when you stack a seminfinite rod of Neodyum magnets. If you make O-shape with slit - you'll get twice of that, up to 1-2 Tesla with the best magnets.
To go further, you would need to compress one magent into the other, hoping that spins would come closer. Normally, that does not work. Though nowone could guarantee you won't get a better magnet by doing this. There are phase transition etc an the whole universe of unknown things. The problem is pressure. We even do not know what is inside our own planet. You migh ptobably need to use neutron stars where gravity helps with pressure. Not in your lab.
Why only 1-2 Tesla? Just calculate B field inside a lattice of, say, iron, putting, say, one spin per atom site....So have you got 1000Tesla? - No. - This is the problem.
> Is my proposal viable?
Try it.
Demagnetisation is relatively simple effect. It comes clear from an excersize of putting two aligned magnetic dipoles aside. One can see that the magetic flux B from first dipole at a point of the secont dipole is directed opposite the field of the second dipole. Therefore, when the nature of dipoles is such that their dipole moment can only be induced by an external filed, and you even know how mouch of magnetic moment you get per unit B-filed for an individual dipole, you still cannot say so easily would it would be if two dipoles are aside, because one acts on another in the aforementioned way reducing the external field and demagnetizing the second dipole.
The main problem is the magnetization, you should use a very high magnetic field in the order of the magnet to be build. Also the problem is that the north pole repels each one (and also south one)
Perhaps if you use magnetized atoms and compress in a z-pinch the M=(1/V)*Sum(mi) said by Marcus would increase a lot the magnetization but you should maintain low temperatures and you should have it some microseconds.
I obtained >10 kiloteslas using a z-pinch without magnets