Magnetic Monopoles? Oh, dear.

Electric charges come in two types, positive and negative. Magnetic poles also come in two types, North and South. In both cases, like charges/poles repel, and opposites attract. The big difference? Electric charges can exist in isolation; you can have just a positive or negative charge by itself. Whereas in a magnet, you always need both a North pole and a South pole; you can’t have a magnetic monopole.

Magnetic monopoles have always been a curiosity for physicists, and many of us think that they ought to exist. In the 1970s, there were searches going on for them, and the most famous one was led by a physicist named Blas Cabrera. He took a long wire and made eight loops out of it, designed to measure magnetic flux through it. If a monopole passed through it, he would get a signal of exactly eight magnetons. But if a standard dipole magnet passed through it, he’d get a signal of +8 followed immediately by one of -8, so he could tell these apart.

So he built this device and waited. Occasionally he’d get one or two magnetons, but the fact that it wasn’t eight was hardcore evidence that something funny was going on with just one or two loops. (Three or more was never seen.) In February of 1982, he didn’t come in on Valentine’s day. When he came back to the office, he surprisingly found that the computer and the device had recorded exactly eight magnetons on February 14th, 1982. Huge devices with larger surface areas and more loops were built, but despite extensive searching, another monopole was never seen. Stephen Weinberg even wrote Blas Cabrera a poem on February 14th, 1983:

Roses are red,
Violets are blue,
It’s time for monopole
Number TWO!

And, as of today, no one has seen good evidence for a second magnetic monopole, leading us to believe that the first one was spurious.

And then I saw this headline over at ScienceDaily:

Damn. And I just tried defending science journalists to a clearly unhappy Sheril Kirshenbaum, and then someone goes and reports, “Magnetic Monopoles Detected…”

You can stop there. They haven’t been. In fact, if you go to the actual paper in science, they say so right in the abstract (bold emphasis is mine, as always):

While sources of magnetic fields–magnetic monopoles–have so far proven elusive as elementary particles, several scenarios have been proposed recently in condensed matter physics of emergent quasiparticles resembling monopoles.

What they did was create magnetic “strings”, or very long, thin magnets on a lattice, where North and South poles are separated by great distances. If you only look at one side of this string, you only see one pole. But the other pole is still there, and so this isn’t a monopole. If you tried to snap the string, you still wouldn’t isolate one magnetic charge; it would work just like this:

Now, this isn’t to belittle the research done here. The creation and isolation of these magnetic strings, and the separation of the poles over long distances is wonderful! As the researchers themselves say, “Above all, it signifies the first time fractionalisation in three dimensions is observed.” So instead of a network of magnetic strings like this:

They’ve managed to separate and isolate a single string! The technique was pretty amazing too, getting down to 0.6 Kelvin to do this experiment. But magnetic monopoles? Please! The science is sensational enough without the addition of outright lying.


  1. #1 Adrianus V
    September 4, 2009

    New Scientist had the same kind of suggestive front, a couple of months ago (see: There story was about the same: condensed matter called spinice has the same characteristics as monopoles have. But no real monopoles in sight.

  2. #2 Blaine
    September 4, 2009

    Thank you so much for clarifying this! When I read the headline, I was pretty excited….but none of you physics blogger types had really said much. I figured if real monopoles had been observed, one of ya’ll would have lit the blogs up! Dammit, this is why I need to finish my damn physics degree, so I can read the papers for myself and maybe figure stuff out on my own.

  3. #3 Stephen Ellestad
    September 4, 2009

    It seemed fishy. Thanks for clearing that up!
    {Shakes fist in general direction of a science journalist} Get off my lawn!

  4. #4 Nomen Nescio
    September 4, 2009

    hypothetical question: if you were to thread such a string through a wire-loop detector such as Cabrera’s, and then when it was partway through, cut the string exactly where it intersected the plane of the loops — what manner of field and field changes would the detector record?

  5. #5 Sili
    September 4, 2009

    Thanks for the full story (and name) for Cabrera. I’ve heard of the failed experiment, but never the details. I think our EM instructor mentioned it second semester at uni – at least that would make sense.

  6. #6 Alex
    September 4, 2009

    This is a cool experiment, and you’re right about the problem with the journalists, but one minor quibble: Is 0.6 K really such an impressive temperature these days? I’m not in that field, so maybe the particular cooling apparatus used is challenging to combine with these particular types of measurements, but it still seems pretty warm by the standards of low temperature physics these days.

  7. #7 NewEnglandBob
    September 4, 2009

    I had a magnetic monopole last week but my cat banged it around the floor until the dog got excited and ate it.

    It was a beautiful cobalt blue one too. Sorry!

  8. #8 'E'
    September 4, 2009

    WHY DON’T YOU REFERENCE THE PAPER ITSELF INSTEAD OF A JOURNAL ARTICLE? Give us some critics on that, if you read it. And for credibility, change your profile picture. It’s not even funny.

  9. #9 Benjohn
    September 4, 2009

    No, you’re wrong “E”, it is a funny picture.

    Ethan, a really good write up, thank you. Nice diagram at the top, and good explanation. It seemed to me (when I read about this earlier) that a string didn’t sound anything like a monopole either, but I figured I’d misunderstood as it was being discussed breathlessly in important journals. Apparently not.

    Ho hum :-)

  10. #10 Benjohn
    September 4, 2009

    Just a thought – any chance that actually, the electrical monopole is the one that doesn’t exist, and we’re always really seeing two electrical dipoles stuck together (+ to + or – to -) by some force we don’t know about? Heh.

  11. #11 Ethan Siegel
    September 4, 2009

    ‘E’, what? The paper is referenced, and referenced heavily. In fact, I link to it and quote from it, in addition to discuss it. I will change my profile picture around November 1st of this year, when I have a suitable replacement.

    Adrianus, although New Scientist was the first magazine to ever pick up one of my scientific papers here, ,

    they also often report on very speculative science, and sensationalize it to smithereens.

    Nomen, let’s say you passed a south pole down through all 8 loops and got your +8 signal. Then you cut the string in the middle of the loops (creating a new north and a new south pole), and pull both ends out. You’ll get -4 from the north pole going down through the bottom four loops, and another -4 from the south pole going up out of the top four loops. So no, there’s no net monopole in your hypothetical situation.

  12. #12 Chrisj
    September 4, 2009

    Sadly, the rubbish doesn’t originate with science daily. It actually seems to have come direct from the Oxford University press office – see for the same errors plus some boggling ignorance. (“So-called because the atomic electrons within it carry ‘spin’…”. Really? You astound me, oh PR-person. What property do normal electrons have instead, then? Non-spinse? [sorry])

  13. #13 'E'
    September 4, 2009

    you quoted the abstract. Not the actual paper. Citing from your same link, but inside the article, you would have read:
    “Our work presents to our knowledge the first direct evidence of Dirac strings. It provides compelling evidence for dissociation of north and south poles—the splitting of the dipole—and the identification of spin ice as the first fractionalised magnet in three dimensions. [...] Our finding [...] initiates the study of a new type of degree of freedom in magnetism, namely an object with both local (pointlike monopole) and extended (tensionless Dirac string)” which is what you were talking about in your blog. The monopoles can be observed at the end of the string. Check out the point where it says “The low energy excitations of such a complex ground state are remarkable in their simplicity, and can be accounted for, to a very good extent, by weakly interacting pointlike quasiparticles, the magnetic monopoles, connected by extended objects, the Dirac strings of reversed spins.”

    I’m just irritated by your statement in the first paragraph in which you affirm “you can’t have a magnetic monopole” when mathematical demonstrations seem to prove differently. This is not the open minded behavior that a scientist should have. I’m not saying not to view it critically, I myself will still re-read the article as there are some points I need to further digest. But in the above blog you didn’t clarify anything, nor proved anything wrong.

  14. #14 F
    September 4, 2009

    E, he’s still correct, magnetic monopoles are still non-existent I think you’re reading to much into the word “can’t” would it help if he changed it to “can’t currently”?

  15. #15 Ethan Siegel
    September 4, 2009

    E, the way these strings work is you have a series of dipoles, NSNSNSNSNS…etc., where the N at one end shows strongly, and the SN are close enough together that they approximately cancel, and the next SN are close enough that they cancel, and so on. But eventually, no matter how you manipulate it in the middle, there’s going to be another S at the other end. This method of playing with the ordering of dipoles on a lattice will never grant you a fundamental magnetic monopole.

    As for your real gripe with me, that I state you can’t have a magnetic monopole, that’s only strictly true given our current understanding of physics. In the standard model, there are no magnetic monopoles, for instance. But extensions, like GUTs (which I am a fan of, if only the proton would decay!), are certainly feasible, and they do have magnetic monopoles. In fact, my PhD supervisor is a big fan of magnetic monopoles. If you google t’Hooft Polyakov magnetic monopoles, his teaching notes come up on the first page. It brings back good memories to think about this.

  16. #16 MadScientist
    September 5, 2009

    Ethan: It’s a pity Sheril doesn’t seem to have revisited her own post and responded to your comments. I’ve only ever dealt with 2 science journalists but they’ve been great; they don’t even rely on what I have to say about something, they go and look up other people to talk to, nor have either of them ever presented kooky stuff for ‘balance’ as some ignoramuses do. The copy editors and others can screw stuff up though, so sometimes the person with their name on a defective article isn’t actually to blame for the problem.

  17. #17 gianmarco bramanti
    September 5, 2009

    Really in the Pauli’s book “Vorlesung uber Elektrodynamik” you can read (I read it in my italian edition): “…main difference with respect to the electric case is the non-existence of isolated magnetic poles. They are always in couple in such a way the total magnetic charge is ever zero. Only in an approximate way we can realize a monopole, building a magnetic needle very long. Gilbert showed that for these monopoles the Coulomb law verify” Pauli book follow this suggestion introducing a very thin analogy between electrostatic and magnetostatic based on magnetic spatial density of charge and surface magnetic density, which really is the divergence of magnetization. Anyway Pauli point out this classic analogy doesn’t overcome the basic difference “there aren’t magnetic conduction charges and currents”. The Pauli’s analogy is based on the correspondence:

    E, H
    B, D
    M, P

    all this is a very simple game. Anyway on the quantum level an other difference was pointed out from the Ahronov Bohm experiment.

    Coming to the Dirac’s work: Dirac introduced ipotetical “conduction magnetic charge” and because he save the minimal coupling and the basic tool of quadripotential, though div(B) isn’t everywhere zero, he write B = rot(A) and this potential vector is singular on a string connected to the monopole. This situation is at all the same in case of the potential vector caused from a Gilbert monopole because outstanding the (needle/string) the induction field is the same. In the Dirac situation, anyway, there isn’t a magnetic induction inside the string that compensate the outcoming flux, so that Dirac exclude the string from the physic domain, and he show this is possible, because a very subtle topological argument, only if the quantization condition is true. In fact Dirac move the string and the charge around in space time, and he can move the string without affect the charge dynamic if and only if the magnetic charge is an integer multiple of a given value.

    Elsewhere the string should affect the charge dynamic and in this case it should be equivalent to a induction field flux tube, though there is none magnetic flux inside, because the Dirac theory postulate the minimal coupling of an abelian gauge theory and the potential vector outside the string is such that rot(A) = B, in other words the quantization condition lacks and the gauge is fixed from the position of the string just in same way it should be, for the causal lienard wiechart choise, if there was a magnetic flux.

    Now the question is very intriguing: are these quasi-particle the pole of measurable string or not. If not these quasi-particle should be very analogoues to the Dirac monopoles charge, elsewhere they should be analogues of the Gilbert’s one, though the Cabrera’s device should distinguish the two situation breaking a real, though invisible (with respect to a fringe experiment) string, pheraps :-)

  18. #18 'E'
    September 5, 2009

    I see. I think I am getting this. But could you please describe more in detail this NSNS cancelation? As far as my understanding goes, I have always considered so far magnetism as a Vector, with the direction in which it is pointing as north and the starting point as south. If in the spin ice these vectors are alternating in the NSNS form, wouldn’t their strength sum up instead of canceling out, just like in normal magnets? The cancelation behavior seems similar to the electric one. Suggesting that indirectly, the string actually contains monopoles, even if these haven’t been separated. Still locally at the end of these strings the effect produced would be analogue to having a monopole. Isn’t it an incredible progress in our technology development anyway?
    Thanks for the clarification

  19. #19 onymous
    September 5, 2009

    “Our work presents to our knowledge the first direct evidence of Dirac strings.”

    That’s odd, because Dirac strings, by the definition I’m familiar with, are inherently unobservable mathematical artifacts. I think they’re (mis-)using the phrase to mean “magnetic flux tube”.

  20. #20 gianmarco bramanti
    September 5, 2009

    If we think these spins just like classic little spinning sphere with their magnetic moment there is no way to obtain any cancellation, which realize in macro-magnets grids of little needles only in an approximate way. Anyway electron in a strong correlated systems are more elusive we can’t imagine, they are relativistic quantum fields. In quantum electrodynamics the fields obey the Maxwell equation, and all the relativistic quantum theory is a consistent gauge theory in which, because the gauge invariance, there’s no room for real magnetic monopoles with exception for the Dirac’s one in a very artificious way, in full agreement with the observation of “onymous” in this scheme we have to consider the strings artifacts and none magnetic flux inside these string is conceived in that case.

    But there’s no conflict with the existence of eventually real strings whit real magnetic flux inside for which the Ahronov integrals vanish, just like the strings doesn’t exists.

    At this level, because the quantum fields are collective systems on which the U(N) group act, we can’t exclude at all that a perfect flux tube at very low temperature, not properly a sum of nano spins, solution of the quantum field equations, is calibrated in such a way the Ahronov integral vanishes.

    If we consider solitons solution for a great non-abelian invariance gauge group, arising in an effective quantum field description of the U(N) thermodynamic limit, then a Polyakov ‘tHooft mechanism may be occurs. Non abelian gauge group really occurs in quantum chromodynamics, where they broke driving to superconductor technicolor condensation inside a quark system, anyway for the solid state superconductors the broken abelian group is what need and suffice, until nowadays application, and should be a very novel and astonishing tool if an electric-magnetic de-facto quantization, vanishing some tipe of gauge integral, occurs.

  21. #21 gianmarco bramanti
    September 6, 2009

    @ “E” A more direct explanation about a way because the string tension vanish along a collective motion wire is in term of exchange interaction generalizing the old Fermi-Dirac pressure of atomic physic. The string tension vanish because an equilibrium is reached, but we can anyway distinguish two terms: magnetic tension (linear in the lenght of string) and Coulomb-Fermi-Dirac repulsion which depend on the details of wave function involved in the collective motion.

  22. #22 'E'
    September 6, 2009

    Thank you for your explanation gianmarco bramanti

  23. #23 Monado, FCD
    September 7, 2009

    I’m no mathematician, but it seems that having one magnetic pole is like having 70 mph without a moving object… or up without down, outside with no inside (Klein bottle, anyone?) or shadows without light; or more precisely, like having a stick with only one end. Can anyone explain without getting into diff EQs?

  24. #24 particle
    September 7, 2009

    I’ve learned a lot from your discussion, my friends. I’m shocked at first when I saw the news. I just couldn’t believe it!

  25. #25 The Quantum Mechanic
    September 8, 2009

    This is an unfair attack on journalists and actually shows how important journalists are.

    The article you cite on ScienceDaily, isn’t a news article, it’s a press release (that’s mostly what ScienceDaily posts). Press releases are not journalism, they are advertising, trying to get journalists to cover the story. Scientific American covered it with the appropriate caveats and New Scientist, which is prone to covering some pretty hokey stuff, only went as far as to say researchers “claim” to have discovered it.

    There is a light-year of difference between a news story and a press release, just as there is such a difference between a product review and an advertisement. To confuse the two is worrying, either because one is unable to tell the difference, or unwilling to tell the difference. If it is the former, than the importance of journalists is clear, to help sift through the information, separating the wheat from the chaff, the signal from the noise. If it is the latter, that is just patently unfair.

  26. #26 SLC
    September 9, 2009

    There is a theoretical difficulty with magnetic monopoles relative to the Dirac theory of the electron. The theory predicts that a magnetic monopole would have attached to it a string of singularities which would appear to violate rotation invariance and thus angular momentum conservation because the string would define a preferred direction in space. Schwinger, in a paper published some 40 years ago, argued that the laws of physics were, in fact, unaffected by the direction the string was pointed in, which had a bit of hand-waving in it.

  27. #27 Gianmarco Bramanti
    September 11, 2009

    It seems just like the rotation invariance violation is permitted only if the quantum magnetic monopole obey to a non trivial spin representation. Because the Dirac duality of Maxwell equation I argue this representation should be better, in view of a more general simmetry theory, if it should be the one half spin representation. A lot of other argument indicate this simmetric role for the elementary magnetic charge: the gauge character of the theory should be verified if the charges obey the fundamental representation of the Lorentz group which is locally isomorphic to the tensor square of SU(2), and the gauge field obey the adjoint representation, which is isomorphic to SO(3)xSO(3). In this way the field angular momentum variation should be permitted to be only a vector one and the quantization condition should be different for a 2 factor in order to have a global invariance of the momentum.

  28. #28 monopole-for-sale
    September 29, 2009

    I have a long solenoid has its ‘N’ end separated from the ‘S’ by one meter, geez, two monopoles just collecting dust…think I can sell them on ebay? buy one get one free? haha.

  29. #29 Ladislav Zobac
    May 22, 2011

    I appreciate very much that I can extend my knowledge of physics by reading about magnetic monopoles, particularly about their detection (or even “production”). I regret that I can not find any description how they look like, except one in a comment here:
    I had a magnetic monopole last week but my cat banged it around the floor until the dog got excited and ate it.
    It was a beautiful cobalt blue one too. Sorry!
    Posted by: NewEnglandBob | September 4, 2009 3:47 PM

    Thank you very much NewEnglandBob, at last I shall be able to recognize it when I meet one!

    But to be serious: Anyone who understands electromagnetic phenomena knows that magnetic field is alway and only the result of moving electrical charges, and “magnetic monopoles” are nothing but words.

  30. #30 Ladislav Zobac
    May 22, 2011

    I appreciate very much that I can extend my knowledge of physics by reading about magnetic monopoles, particularly about their detection (or even “production”). I regret that I can not find any description how they look like, except one in a comment here:
    I had a magnetic monopole last week but my cat banged it around the floor until the dog got excited and ate it.
    It was a beautiful cobalt blue one too. Sorry!
    Posted by: NewEnglandBob | September 4, 2009 3:47 PM

    Thank you very much NewEnglandBob, at last I shall be able to recognize it when I meet one!

    But to be serious: Anyone who understands electromagnetic phenomena knows that magnetic field is alway and only the result of moving electrical charges, and “magnetic monopoles” are nothing but words.