If I place many micro-bar-magnets on the surface of hollow sphere such that the pole of every magnet points towards the same side then, you could imagine that a monopole would be produced. So, I just created a monopole. But we all know that monopoles don't exist. Then what is wrong in this procedure? Assuming that I got some kind of power that I can put the magnets together in one place without getting budged by the repulsive forces!


A good starting point to see that gluing small magnets to a sphere doesn't work is seeing what the field of just two magnets opposed gives you. A single dipole magnet has a field that falls off as $\frac{1}{r^3}$, while the two opposed has a field that dies faster, falling off as $\frac{1}{r^4}$. As you add more and more magnets (pointing in different directions), you cancel out successively more and more of the field (i.e. higher order moments), and eventually, in the limit of infinite tiny magnets all pointed outward, the sum of all the magnetic fields cancel, and you are left with, oddly enough, no magnetic field at all, inside or outside the sphere (except for inside the magnets themselves).

Edit: A good way to think about this is considering a magnetic dipole as two magnetic monopoles very close to one another. So this sphere of magnets can be thought of as two spheres of uniform magnetic charge and opposite sign. By the shell theorem, there is no field inside a uniformly charged shell, and outside the uniformly charged shell the field is (in some system of natural units) $\frac{Q}{r^2}-\frac{Q}{r^2}$. The only place the field is non-zero is between the shells. If we take $r_{out}-r_{in} \rightarrow 0$ while keeping $\frac{Q}{r_{out}-r_{in}}$ constant (which keeps the dipole moment of each dipole constant), this field vanishes as well.

Yet another edit: I was asked for more clarification, so I'll go about this another way:

First, it's clear that whatever field we find must be rotationally symmetric, as the system we have set up has spherical symmetry. This forbids the existence of any fields (using spherical coordinates) in the $\hat\phi$ or $\hat\theta$ directions (azimuthal or polar directions). So the only field we can have is radial, and cannot depend on angle: $$\vec B=f(r)\hat r$$

Now, we can apply Gauss' law for magnetic fields (with magnetic monopoles):

$$\int \vec B \cdot d\vec A = \mu_0Q_m$$

where the left hand side is a surface integral over an enclosed volume, and $Q_m$ is the amount of magnetic charge contained within. If we use a spherical volume, then the left hand side simplifies: (as $\vec B \mathbin{\|} \vec dA$ and $\left|\vec B\right|$ is constant)

$$\int \vec B \cdot d\vec A = \int\left|\vec B\right| \left|d \vec A\right| =\left|\vec B\right|\int\left|d\vec A\right| = \left|\vec B\right|4\pi r^2$$

And thus the total magnetic field is:

$$\vec B=\frac{\mu_0Q_m}{4\pi r^2}\hat r$$

In the absence of any magnetic monopoles, $Q_m=0$, and so this is zero. If you make this and notice a monopole moment, it means that there is a magnetic monopole that just happened to stop inside your sphere! Don't lose it- it's probably worth a Nobel prize.

  • $\begingroup$ You mean inside the sphere? $\endgroup$ – anna v Feb 13 '16 at 12:31
  • $\begingroup$ Inside or outside the sphere, actually. Except for inside the magnets themselves. I'll clarify. $\endgroup$ – Chris Feb 13 '16 at 12:32
  • $\begingroup$ I do not think this can be correct,. you are talking of dipoles in series when you give the functional form, not on a sphere $\endgroup$ – anna v Feb 13 '16 at 12:36
  • $\begingroup$ @annav See my edit, hopefully this clarifies it some more. It is a very strange phenomenon to see for the first time. $\endgroup$ – Chris Feb 13 '16 at 12:38
  • $\begingroup$ I would like to do the experiment. I can see , as I said in my answer, that the return lines will be bunched and go through the sides back to the north pole, but the strength of the repulsion, south south inside, will be great. You are saying that because of the bunching of the return field lines, the north field at the limit disappears too, (is absorbed in the material) whereas I see great repulsion for the inside poles that will not allow for a forming of a tight sphere. But you are talking of reducing the size of the dipole. I am thinking of a real experiment. $\endgroup$ – anna v Feb 13 '16 at 12:47

One can construct such a shell, if one really wanted to, by many individual magnets attached on a spherical shell, almost parallel to each other north poles out south in, for example.

With run of the mill magnets, there will always be enough space between them so that the lines will return from the sides to the outside pole. The repulsion will be very great, i.e to hold the magnets in place will need strong ties. The field lines would all bunch back on the sides to go towards the other pole, and inside the field would be zero, certainly at the center.

I suspect if one designed well fitting magnets, so no space was left between them , even of micron dimensions, there would be an enormous repulsive force and the magnet could not be held in place.

  • 1
    $\begingroup$ Hmm.. talking mathematically .. We are considering small magnets.. like really small. So the magnetic monopole +M and -M would be small too, wouldn't it? So I guess there is a possibility of getting a finite limit of 0/0 form or not? $\endgroup$ – brainst Feb 13 '16 at 14:05
  • $\begingroup$ @brainst May be you are? The question itself is not about micro magnets. Of course you can take a limit, but I am answering the question as is. $\endgroup$ – anna v Feb 13 '16 at 14:11
  • $\begingroup$ You don't need space between the magnets for the field lines to loop around. The field lines from North Dakota will loop back through South Dakota and Minnesota.....cancelling out the fields from those neighboring states. That's why there is no net field and you can't construct a magnetic monopole. $\endgroup$ – Marty Green Feb 13 '16 at 14:31
  • $\begingroup$ @MartyGreen Well, field lines drawn for ferromagnets are not shown through the material. Of course the circuit closes, but it will need a separate study if it closes through the iron, in the experimental set up I am envisaging, because instead of unidirectional in the material it will have to be bidirectional, imo $\endgroup$ – anna v Feb 13 '16 at 15:43
  • 1
    $\begingroup$ @MartyGreen an experiment always succeeds, if the measurements are correct. You mean if an experiment validates my hypothesis , I suppose. But my hypothesis is not so far off from all the others, it is just macroscopic, not taken to a limit And it will not be a monopole. Make an electromagnet a trumpet like solenoid, and make the wide side very extensive, so the field lines will be very diffuse at the wide end , and point like at the narrow, will it be a monopole at the narrow? $\endgroup$ – anna v Feb 13 '16 at 17:03

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