How do Magnets Work?
Most recent answer: 10/22/2007
- Marsha Borski (age 9)
That’s an excellent question.
Try the following experiment with a magnet, some needles (the plain kind used to hold stuff together as you sew it), and a screwdriver. It will help you understand the explanation.
Touch a needle with a screwdriver and nothing unusual should happen (unless the screwdriver is already magnetized). Now, take some of the needles and hold them against the end of a magnet for a while (the stronger the better). After doing this, you will have turned each needle into a little magnet, and when you touch them with a screwdriver they will stick to it.
Some kinds of metals (like steel that the needles are made of) are made up of billions and billions of individual atoms that each have the properties of a microscopic magnet. The atoms in steel naturally tend to get together in tiny little groups called domains, and within each domain the atoms tend to point in the same direction, which makes the domains behave like a tiny little bar magnets just like the kind you have probably played with at school. The needle of a compass is also a bar magnet, and we know what this does: it points north because it likes to line itself up with the magnetic field of the earth.
When you make a piece of steel (like a needle), all of the tiny domain-magnets inside tend to get stuck pointing in different directions, which means that they more or less cancel each-other out, so to begin with the needle does not behave much like a magnet at all.
However, if you bring the needle close to another magnet of some kind, (like the ones on your fridge, or better yet some stronger ones you can find in your classroom) something interesting can happen: Because of the other magnet you are holding it near, the little domain magnets in your needle will tend to line up so that they are pointing along the same direction. (You can visualize this by holding two bar magnets near each other and noticing how they like to line up a certain way). The domains in the needle will do the same thing, and after you take the needle away from the other magnet, the domains in the needle will tend to stay pointing this way (they sort of get stuck pointing in the same direction). Now, since you have lots of these little domain magnets in the needle all pointing in the same direction, the needle will itself have become a small magnet. (This is exactly how bar magnets are made).
Magnets attract other magnets (as you can see), and also attract some kinds of metal like many kinds of steel. If you touch your magnetic needle with a steel screwdriver it will stick. However, stainless steel is not a very good magnetic material, so if you touch your magnetic needle with something made of stainless steel it will probably not stick (try it).
(published on 10/22/2007)
Follow-Up #1: why not more magnets?
In most materials almost all the electrons form pairs, with the magnetism from the two paired electrons exactly canceling. The result ultimately is due to something called the Pauli exclusion principle, which says that no more than one electron can exist in any particular quantum state. So if there's some nice low-energy state waveform for an electron to sit in in some molecule, it tends to get two electrons for the two possible quantum states with that form: one spin up, the other spin down.
This very fact is the origin of most of the properties of chemical bonding. Remember how a typical covalent bond consists of two shared electrons? The reason is the exclusion principle plus spin. In ionic bonds, typically one electron will mostly move from one atom with an odd number of electrons over to another atom with an odd number, leaving each with an even number.
So the tendency of most materials not to be magnetic is closely tied to the same quantum mechanical facts which account for the prevalence of pair-bonds and ionic bonds in chemistry.
(published on 10/22/2007)
Follow-Up #2: sad truth
- ivan brown (age 13)
Newport, Wales, Britain
(published on 10/22/2007)
Follow-Up #3: non-magnetic trees
- lai (age 18)
In leaves, however, as photosynthesis occurs some electrons get shuttled around individually as part of the light-driven chemical process. So their magnetism isn't automatically canceled, and magnetic resonance signals from them can be measured.
(published on 10/22/2007)
Follow-Up #4: pauli exclusion principle
- starlai (age 18)
(published on 10/22/2007)
Follow-Up #5: domain interactions
Think of it this way. If you have two bar magnets, each has a North pole and a South pole.
If the two magnets have their two N poles pointing up then when they are side by side they will repel each other but when they are head-to-tail they will attract. In the other case when one has a N pole pointing up and the other has a S pole pointing up then they will attract each other when they are side by side but repel when they are head-to-tail.
(published on 12/17/2007)
Follow-Up #6: permanent magnets
- Anonymous (age 13)
To make a permanent magnet, you need some way to keep those domains lined up even after the applied field is removed. In certain special materials, it's hard for the domains to change which way they point. Usually those materials have some mixture of non-magnetic atoms, which tend to catch and trap the domain walls, the regions where the magnetism changes from one direction to another. Trapping the walls keeps them from moving, and keeps the magnetic direction from changing.
If you heat up a permanent magnet, the walls tend to come loose and the net magnetism will go away.
(published on 02/19/2008)
Follow-Up #7: magnetic vs. electric fields
One of the sources of magnetism is electrical current. That's why a current through a copper coil, for example, creates a magnetic field. Now (and this isn't quite being honest) imagine a charged particle as a little ball of charged stuff. If it's spinning, then that's like having currents go around in a loop, and you should get fields like those when current goes around in a coil. Elementary particles like electrons do indeed have "spin", measurable angular momentum. However, if you took this story too literally you'd calculate a magnetic moment of an electron which is only half the actual value. On a small scale, that spin is really something quantum mechanical, not something you can picture as charges running around in circles.
Magnetism is related to the electrical fields which pull on charged particles. However, it's connected not just with the positions of the particles but also their velocities. If you look at a charged particle (without spin) from a frame where you say it's standing still, it will only have electrical fields. If you look from another point of view and say it's moving, it will also have magnetic fields. Of course the net physical effects on its neighbors will be the same in either case. We obviously cannot fully introduce this area here, but if you want to see a beautiful book on the topic, I recommend E. Purcell's "Electricity and Magnetism"..
To add one more twist to the tale, Maxwell's equations are consistent with terms involving magnetic monopoles. None have ever been seen. For more information check out:
And for another twist, if those magnetic monopoles exist, then an observer who sees one moving will say it produces electric as well as magnetic fields. mike w
(published on 02/24/2008)
Follow-Up #8: making magnets and demagnetizing them
So you're absolutely right that an ac current is no way to magnetize something. In fact, when you want to demagnetize something (e.g. erase a magnetic recording) the standard way is by applying a big ac field and gradually decreasing the magnitude. That pretty much scrambles up the domains. It's often called "de-Gaussing" since one standard unit of magnetic strength is a Gauss.
(published on 03/08/2008)
Follow-Up #9: Magnets do not bend light, however...
- Joe Shock-Ra
There is a subtle effect, however, when light passes through a dielectric medium like glass or sugar water solution. It turns out that in the presence of a magnetic field the light beam is not bent but the plane of polarization can be rotated a bit. This is called Faraday rotation. You can find out more about it at
(published on 04/25/2008)
Follow-Up #10: magnetic forces
- Tiffany (age 13)
Richmond hill, GA, US
Think of the gravity that holds you and the Earth stuck together. As long as you and the Earth are around, that gravity will be there. The same rule applies to magnets. As long as they stay magnets, the force between them will be there. There is one difference, however. You can think of a magnet as made up of little magnets, called domains. Each one has a North and a South pole. Over a very long time, those can jiggle around so the little magnets don't line up with each other. Then they won't add up to make a big magnet. Gravity doesn't have poles, the gravity from every part just adds together. So even as parts of the Earth shift around, the gravity won't go away.
(published on 03/30/2009)
Follow-Up #11: compass physics
- John (age 33)
The inaccuracy near the north pole arises because the magnetic north pole isn't actually at the same place as the rotational north pole. The difference becomes important in that vicinity. If you stood in between them, your compass would point south.
(published on 09/23/2009)
Follow-Up #12: Do two bar magnets attract or repel each other?
- Drew (age 17)
| | | | Currents aligned
| N A S | | N B S | Attraction
| | | | Currents anti-aligned
| N C S | | S D N | Repulsion
Just to be clear, in the picture above A and B attract, C and D repel. You might also wonder about what happens between A and C and between B and D, in the positions shown. A and C repel. B and D attract. If you imagine each magnet as a tube of current going around the axis, the ones that attract are the ones where the nearby parts of the current go the same way in each tube. Mike W.
(published on 03/24/2010)
Follow-Up #13: magnets attract or repel
- Isabel (age 10)
Bar magnets only sometimes stick together and sometimes repel because each has a north pole and a south pole. Opposite poles attract, and like poles repel. If you were to put two north poles together, they would repel, but a north and a south would stick together. The picture above shows an example.
Chris F. and Mike W.
(published on 02/16/2011)
Follow-Up #14: are magnets permanent?
You also ask, if I understand right, what sort of energy is involved in the magnetic forces between magnets. There is energy stored in the magnetic field itself. The density of that energy is proportional to the square of the field strength. When magnets move near each other, that field energy generally changes.
(published on 07/04/2011)
Follow-Up #15: shielding a magnet with steel
- SARAH (age 26)
Here's a nice physical way to picture what happens. Picture the magnetic field in the usual way, as a set of field lines coming out of one pole of the magnet and returning to the other. There's a tendency for those field lines to get sucked into pieces of iron or steel or other easily magnetized material. That's exactly the reason that the paperclips stick to the magnet. Some of the field lines get stuck in them.
Now what happens if you put a big sheet of steel in between the magnet and the paper clips? The field lines mostly get drawn into the sheet, spread out in the sheet, and return without going past the sheet to the paperclip. So the clip doesn't stick much if it's just behind the sheet.
This effect depends a lot on the shape of the sheet. A small piece of steel can pull field lines into it leaving more field just behind it than were there before. That's just what happens when you chain paperclips together hanging from a magnet. Each clip tends to steer field lines on to the next one. So the steel redirects the field, with a sheet redistributing it away from the clip just on the other side. It would tend to grab clips right near the edge of the sheet, where some of the field lines exit.
(published on 09/28/2011)
Follow-Up #16: magnetized nail
- Karen (age 10)
When you put the nail in the electromagnet made with the coiled wire powered by the battery, it lines up a lot of those domains to point the same way. After you take the nail away, some of the domains stay stuck, so you're still left with more of the magnetism pointing one way than any other. That means the nail is a bit magnetized.
This is temporarily posted without the usual check, until Lee gets back from Paris.
(published on 11/27/2011)
Follow-Up #17: magnets and their fields
- Atticus (age 14)
Salt Lake City, Utah,
Sure, you can make a donut-shaped magnet with the N poles on the outer part and the S poles near the donut hole. It's not a standard form, but it's certainly possible. (Those poles aren't really "charges" though.)
Your second question is extremely hard to answer. I think for now perhaps the best non-answer is this. Whatever the universe consists of, at some point our description has to get down to the most basic ingredients, or, if every level of ingredient should be thought of as made of something deeper, at some point we'll just get to the deepest level we know about. At that point, all you can do is describe the properties, not say what things are made of. Electromagnetic fields are pretty close to that point. We could give maybe a step or so deeper description, but it would be in terms of quantum fields. Those are even more abstract mathematical entities than magnetic fields.
(published on 01/05/2012)
Follow-Up #18: magnetism from moving charges
- Rahul (age 15)
1. Start with the existence of simple electrostatic forces following Coulomb's law.
2. Assume that special relativity gives the right rules for how things look in different reference frames.
3. Look at a moving charge near a neutral, current-carrying wire. There's no electrical force.
4. But if you now look in the rest frame of the charge, the wire is no longer neutral, thanks to the different Lorentz contractions of the differently-moving plus and minus charges.
5. So there's an electrical force in this frame.
6. So there must have been a velocity-dependent force back in the frame where the wire was neutral.
7. We call that velocity-dependent force magnetism.
Ok, that handles the part about why magnetism, given electricity. The part about why electricity in the first place is unfortunately over my head. Here's a few words to get you started. Purcell derives the magnetism from electricity from assuming a symmetry, special relativity, a rule about how the world obeys the same laws of physics even as you represent its contents in different ways. There are some other subtle symmetries (gauge symmetries) in relativistic quantum field theory, and I've heard that they require the electrostatic force.
(published on 05/28/2012)
Follow-Up #19: first magnets
Really nice questions!
Some stones are magnetic as found. They are magnetized because they cooled in the Earth's magnetic field. The Earth is magnetic for complicated reasons which I don't really understand, but have something to do with currents stirred up as heat flows from the hot core to the cooler surface, creating convection patterns.
You don't need a permanent magnet to magnetize a piece of iron. It's usually done using an electromagnet, so what you really need is a current source to drive the electromagnet. Although many current sources are ultimately based on generators which themselves use magnets, there are other current sources which do not need magnets. For example batteries can be made using chemicals and no magnets.
(published on 05/16/2013)
Follow-Up #20: magnetic water?
(published on 05/16/2013)
Follow-Up #21: how do magnets act at a distance
- Steve (age 28)
We were just discussing doing more to encourage our readers to do experiments. Your suggestion about raising follow-up questions for them is along the same lines. We'll do our best.
On "action-at-a-distance", we don't really think of it that way. We think of the magnetic field as a real thing, all spread out in space, acting where it is. But how did that field get all spread-out? Not by sudden action-at-a-distance. It had to work its way over from the source, by an electromagnetic wave traveling at the speed of light. That's not much of a coincidence, since light itself is an electromagnetic wave.
We don't really know what anything is made of at the deepest level. I guess the two basic approaches are that everything is some sort of differential equation (a little bit like Maxwell's equations for electromagnetism) or that everything is some sort of cellular automaton, kind of like a bunch of little digital bits. Right now, the deepest things we know are all differential equations, but some people suspect that may change.
Saying "It's all just organized energy" doesn't really convey a definite enough idea for us to say it's true or false.
(published on 10/03/2013)
Follow-Up #22: Adding bar magnets together
- Mark Basham (age 50)
If you have two thin bar magnets and put them together in parallel the resulting strength is about double. If you keep adding more and more of them eventually the resulting sum is not the sum of the number of magnets due to the over all geometry of the combination. It involves some integrals.
If you add them lengthwise you don't get a factor of two in the end-on direction but you would get a factor of two if you placed them face down on an iron surface.
Far away, you do get a factor of two in the fields, even though close by it's more complicated. /mw
(published on 12/19/2013)
Follow-Up #23: Does positive attract positive?
- Moe (age 16)
Positive energy does attract positive energy by ordinary gravity. However positive electric charge repels positive electric charge. This old physics has nothing to do with various new-age ideas.
Magnets are still another matter. No magnetic charges ("monopoles") have ever been found. However, like poles of magnets (e.g. North and North) repel.
Your mom may not be nuts because many medical conditions are known to be helped by something called the placebo effect. People who believe that something (e.g. an alleged medicine) will help them tend to do better, sometimes in real objective measured traits, not always just in how they feel. () Magnets are truly harmless. Maybe having to pay a lot for them made her feel like they're doing more good, and maybe that feeling makes them do more good. I have a lot of trouble motivating myself to exercise enough. Maybe it's time to invest in some magnets.
(published on 02/04/2014)
Follow-Up #24: Storing energy in magnets
- Anonymous (age 15)
As you point out, you can't create energy with magnets, but you can store energy. When you push two same-sign poles together it requires energy. This energy can then be released and do useful work when you let go of the magnets.
(published on 03/17/2014)
Follow-Up #25: natural magnetism
- Juan Preciado-Riestra (age 42)
Yes, ferromagnetic materials typically do have a weak net magnetization. The Earth is a weak magnet (that's why compasses work) so the ferromagnetic material will magnetize a little as it cools into the magnetic state.
(published on 04/27/2015)
Follow-Up #26: what is a magnetic field?
- Dennis (age 45)
1) You could always remove the energy for use once, just by letting two oppositely polarized magnets attract and do work as they approach each other. That's just like how you can get work out of letting some object fall in a gravitational field or from the attraction of opposite electric charges. Once it's gone,though, it's gone. There are no perpetual motion machines.
2) Just that fact that you have to heat the magnet tells you you're mostly adding energy. That's because the magnetized region has negative interaction energy between its spins. It also has a little positive field energy. As the material demagnetizes by individual spins flipping over, each emits an electromagnetic wave. A little energy can leave the material that way, if it's not absorbed and turned into general thermal shaking of parts.
3) See this: .
5) The spins within each domain are mostly aligned with each other. The field isn't going anywhere.
7) When the field was first set up, energy did have to flow there. The flow density is described by the Poynting vector, proportional to E X B. So during that process there was also an electrical field. You wouldn't see the reason for that if you came along after the process was over and everything was static.
8) How does the Sun's gravitational field extend out to us? These fields propagate very fast, at the speed of light. Your question seems to assume that things are made of little parts that interact only when they touch each other. Michael Faraday struggled with pictures like that before working his way over to something like our modern idea of fields.
9) I suppose there's a theoretical maximum at which the field energy is enough to collapse the object into a black hole. Maybe there's some other theoretical maximum that's not so huge, but I'm not aware of it. For the types of magnets we're familiar with on Earth, the practical maximum would have a few spins, and thus a few Bohr magnetons, for each atom. That magnetization corresponds to about 10,000 Gauss, which is around what the most intense permanent magnets reach. Big superconducting electromagnets can reach over 100,000 Gauss. Higher values can be reached briefly. For very different types of matter (neutron stars) fields can reach over 1014 Gauss. () At that point the field energy density is pretty significant, about a million times higher than the rest mass energy density of water. That's still a lot less than the energy density of the magnetic neutron star itself.
(published on 07/15/2015)
Follow-Up #27: magnetic domains
- Kevin Jin (age 15)
First, on the "line up" issue, you raise a good point. Quantum mechanics does not allow the spin to be fully lined up in any direction. If the average spin is in the z direction, there will be an uncertain spread of possible spins in the xy plane. Still, that allows the average values of different spins to line up with each other, despite the leftover spread.
The domains form because the electron waves can lower their energy when nearby spins line up. That's not a general property of all materials but rather depends on the quantum states of the electrons in the atoms, on the types of neighbors, and on the distances between the neighbors. In iron, for example, different densities of different amorphous alloy forms can either favor alignment or anti-alignment.
So if the material happens to be one in which alignment is favored, why doesn't it form one big domain? As the domains get bigger, the plain old magnetic interaction, like you see when you play with magnets, makes adjacent regions want to anti-align if they are sitting lateral to the field direction. Ones displaced along the field direction, in contrast, lower their magnetic field energy by lining up the same way. Thus ordinary magnetic material is a patchwork of domains pointing different directions. A thin needle, however, can form a single domain with the field along the needle direction.
(published on 10/03/2016)
Follow-Up #28: magnetic direction
- Conner (age 16)
San Diego CA
Quite aside from the shape of the electron cloud, the electron itself has an intrinsic property called "spin". That gives it some angular momentum and magnetic moment, which point in some direction. So even a nice simple hydrogen atom with a spherical electron cloud already has magnetism with a direction.
(published on 11/15/2016)
Follow-Up #29: electron spin in magnets
- Freddy (age 24)
The electrons in the north and south poles of a bar magnet are spinning the same way, not opposite ways.
(published on 02/18/2017)
Follow-Up #30: magnet repel or attract?
- Freddy (age 24)
Hi Fred- Unfortunately pictures don't come through in our question sysem, which has some filters designed to stop various hacks. The internet has turned into a dangerous place, and that interferes with ordinarye life.
I'll try to reconstruct what I think you're asking. Spin directions will represented by little arrowheads.
N<<<<<<<<< S S>>>>>>>>>>N. These two repel.
Flip the second around the middle of its long direction and you get:
N<<<<<<<<< S N<<<<<<<<< S These two attract.
But maybe you were asking about
N<< S S>>N
N<< S S>>N
N<< S S>>N
N<< S S>>N
N<< S S>>N
These two repel.
Flip the second around the middle of its long direction and nothing changes, they still repel. Perhaps that's the picture you were describing.
Flip it around the middle of its skinny direction and you get:
N<< S N<< S
N<< S N<< S
N<< S N<< S
â€‹N<< S N<< S
N<< S N<< S
These now attract.
I suspect that you took the spins from one picture and the N,S labels from the other.
(published on 02/22/2017)