The Universe in a Bottle

Dr. Robert B. Sheldon
To see a World in a Grain of Sand
And a Heaven in a Wild Flower,
Hold Infinity in the palm of your hand
And Eternity in an hour.
William Blake

Space Plasma Physics is the study of the most common form of matter and the majority of the Universe. Plasmas are the "fourth" state of matter, where the first is cold solid, then warm liquid, cool gasses, and finally, very hot plasmas. Stars, nebulae, quasars, are mostly plasma. Plasmas can be millions of degrees hot, and yet like the stuff inside a fluorescent light tube, still not melt the walls just because it is so very tenuous. Solids and liquids have densities of grams/cubic-centimeter, and gasses are grams per 22 liters, but plasmas are grams per million cubic kilometers. The only reason we know they are there, is that as we look through a million miles of the stuff, we can detect its faint glow. Or, more often, it heats up the gas around it, and the gas glows. In the pictures below, that's how we manage to get pictures of the plasma.

(That's Geoff, a physics major, standing next to the apparatus he build for about $4000 with used parts off e-bay.)

Well, why isn't plasma just hot gas then, how do we tell the difference? Very simple actually, wave a magnet nearby, and gas won't budge but the plasma will swirl and dance. Not because plasmas are magnetic--though they can carry enough electrical current to become electromagnets--but because electricity is bent by magnets, and plasmas are just congealed electricity, equal numbers of negative electrons and positive ions. And that is the secret to getting the universe in a bottle.

You see, because plasmas are so very hot, too much gas will cool them off. So in order to make them in the lab, we have to pump most of the air out of the light-bulb or bell jar, and then heat the remainder with high voltage. And because the plasma is so hot, it wants to expand but then it hits the walls where it cools and disappears. Plasma physicists call this "wall effects", and it means that the little bit of plasma captured in our lightbulb looks very different than the wild plasmas out in the huge vacuum of space. Space plasma physicists then have to plead for spacecraft safaris to travel to outer space and record their native behavior, or, more cheaply, model them in their computers at home.

But magnets turn out to be a plasma cage without bars. Because a magnet bends the plasma, we can put a magnet in the middle of a plasma, and the plasma will swirl around it like moths around a candle. And the stronger the magnet, the more closely the plasma swirls around it. In the past, magnets made of iron or AlNiCo could only keep the plasma within a few feet, so plasma experiments had to have huge vacuum jars--a meter or two in diameter. Today, the new Neodymium-Iron-Boron (NIB) magnets are ten to a hundred times stronger, and a few centimeters can suffice to contain the plasma. This is fortuitous, because glass bell jars come in 30 or 40 centimeter sizes that are very convenient to pump down and make plasmas.

A cylindrical NIB magnet was mounted on a hollow ceramic straw (any metal or plastic will do), with a wire threaded up the middle, and put in a bell jar. The bell jar was evacuated with cheap oil-piston vacuum pump (HVAC techs use it to refurbish your air conditioner), to a pressure of about 100mTorr, which is the best they can do, but perfect for us. This vacuum is called "The Paschen Point", because it is the pressure at which gas will glow when you put a high "breakdown" voltage on it--like the inside of neon or fluorescent bulbs. And even more amazing, the magnet lowers the breakdown voltage from a few 1000VAC down to -400VDC. The electrons boil off the magnet, knocking other electrons off some of the air molecules in the bell jar. Those air molecules, now "ions", are suddenly plasma and get trapped by the magnetic field.

Magnetic Storms

My first try involved putting a sharp point behind the magnet to make a glow discharge plasma that would get trapped by the magnet. You can see the glow to the right of the magnet where the plasma was made, and just a tiny bit of it circulating around to the left side of the magnet. But what I didn't expect where the arcs, the little points of light on top of the magnet that look like lightning flashes striking the magnet. A bit further down on this page, we attribute this to electrons being pulled out of the plasma, but I spent a week trying to figure out what they were. You will notice there are more of them on the right side where the plasma is being made. When I performed this experiment, I had hoped it would explain how plasmas injected in the Earth's tail could extract ions from the Earth's atmosphere, and this is what those points illuminate--an electric field following a magnetic field line down to the magnet (representing the Earth).

The truth of the matter is that I never expected to have wild plasmas in a jar; I thought I was going to be experimenting with the tame variety found in lightbulbs. But when I saw the donut-shaped plasma swirling around the NIB magnet, in a flash I knew I was seeing the shape of the Earth's radiation belts--the dangerous space plasma encountered by the very first US satellites. From then on, I have been able to capture many other space plasmas, and discover things that the computer jockeys never thought to look for.

So these are more than pretty pictures, they are examples of things we have only viewed from afar or sampled poorly with our spacecraft. They are solutions to magnetohydrodynamics equations for which computers have given us "partial" and "approximate" answers. But most of all, they reveal the beauty, the simplicity, and the majesty of the universe and the one who made it.

Earth's Radiation Belts

In this series of pictures, we slowly change the pressure in the bell jar. Right bottom at high pressure (but still a vacuum!), you can see the blue glow of electrons as they hit the air, and the thick schmere of pink ions colliding and bending around the equator of the magnet. The nitrogen molecules that are hit by ions glow pink, but for our 8-bit camera, the color saturates to white. At low pressure (better vacuum) counter-clockwise, you stop seeing the electron glow because its too faint, but the ions are now travelling in circles around the magnet without smearing out (diffusing), so they get more pancake-shaped. While the Earth's radiation belts have slightly different origins (cosmic rays hitting the atmosphere plus electrons heated in the cusps), they are trapped exactly the same way, and look exactly the same.

Earth's Quadrupole Cusps

I mentioned that the electrons in the Earth's radiation belts come from heating in the cusps. A cusp forms whenever two magnets are brought close together. We can bring them together one of two different ways--either so that the magnets repel or attract each other. When they repel, then there are two spots where the magnetic field from each of the magnets is exactly equal and opposite, and so the "strength" of the field in that spot is zero. We call that zero field spot a "cusp" because of the way it looks.

But where is the 2nd magnet to make cusps at the Earth? It turns out it is the Sun, and here's how it works: Nature always likes to keep things stable (a principle known as Lenz' law), so when we bring a strong magnet up to a sheet of copper, it repels. Why? Because the electrons in the copper get swirled in circles by the approaching magnet, which creates an electromagnet in the copper that repels the NIB magnet. All materials behave this "diamagnetic" way, copper more obviously because it is such a good conductor that the electrons really get moving. (Your refrigerator behaves differently than copper, because steel has a very unique property due to a nefarious collaboration of iron atoms called "ferromagnetism" that overcome the diamagnetic repulsion.) But plasma has even more liberated electrons than copper, acting like a superconductor. And plasma is what the Sun is blowing at the Earth all day long. That means the Earth feels like the Sun is pushing a magnet up close to it, and the Earth pushes back, so the Sun's "solar wind" has to go around the Earth. But there's always more wind arriving from the Sun, so a "steady state" is reached where the Earth's magnetic field is all wrapped up in this "wind sock" shape that we call unimaginatively "The Magnetosphere".

In the laboratory, we can simulate most of this geometry by making a radiation belt around the 1st magnet, which is accomplished by putting -400VDC on the magnet. Then we bring up a 2nd magnet but we don't put any voltage on it. So the plasma created on the 1st magnet is pushing up against the magnetic field of the 2nd magnet, and we discover two things by looking at the plasma glow. First, the plasma doesn't cross over a line (known as the separatrix--yeah, Latin). And secondly, the only place the plasma seems to cross is right at the cusps, those two "horns" that come out near the top and the bottom of the 2nd magnet.

Why there? Because that's the only spot where the magnetic field strength goes to zero, so it doesn't repel the plasma. Why do the cusps light up? Because they are good traps for the plasma as well.

And now you've learned two things that very few computer jockeys knew or NASA safaris realized.

Earth's Aurora

The aurora is made by electrons (and very occasionally ions) that cause the atmosphere to glow. The power source for those electrons is the solar wind, whose constant blowing creates a voltage across the magnetosphere's wind-sock "tail". When electrons feel this voltage, it gives them energy and a kick to travel back toward to the Earth, where they slam into the atmosphere like suicide bombers. This rain of electrons makes the beautiful shimmering curtain that we call the aorora borealis. (Okay, there's also an aurora australis, but unless you overwinter in Antarctica, you're unlikely to see it.)

We can simulate this in the lab by setting up two magnets just like we did for the Earth's cusps. But this time we make the 2nd magnet positive +5VDC. This isn't high enough to make any plasma, but it does pull electrons from the 1st magnet right down on top of itself. Now as the electrons get closer to the magnet, they generate that electromagnet we talked about earlier, the "diamagnetism" that repels them. That's why we had to crank up the voltage from 5, to 10, to 15 VDC to overcome that natural repulsion. And when we did that, we see the same "curtain of light" appearing over the top of the magnet.

So what happens when you turn the voltage up higher? Well it gets brighter and closer to the magnet until there's this explosion, and the power supply circuit breaker cut out. It was all over too quickly for my 20 frames/second movie camera to catch it, but what little I saw looks just like a northern lights display known as a "substorm". The "C-shaped" northern lights develop a bright spot, and the bright spot spreads sideways until the entire auroral oval just sort of lights up with this explosion. I think in the lab, the electrons start hitting the magnet, which knock off ions, but the ions coming up knock off electrons from the gas in the chamber, and those electrons go back down and hit the magnet, and the whole thing exponentially draws more and more current until the power supply trips.

And now you know more about substorms than all the antagonists in the "substorm wars" learned in 30 years.

Reconnection X-line

We have done two experiments with the two magnets arranged in the "repelling" mode. What happens when we flip one magnet over and look at the "attracting" mode of two dipoles? That's what this picture shows. But what does it mean?

I'd have to draw a picture of magnet with little arrows coming out of it to prove it, but if I did, then the place where the two magnets neutralize each other is halfway between the two of them. But it isn't just a single spot, like it was when the magnets repel, rather a whole "line" or volume of magnetic fields cancel each other. This means the plasma on one magnet doesn't experience this "separatrix" that keeps it apart except for the tiny hole, but the plasma on one magnet can easily jump over to the other magnet.

But remember that magnets make the plasma move sideways. So what happens when the plasma jumps from one magnet to the other? Think of it as two egg beaters whipping up some lemon meringue. Each egg beater represents a magnet, and the meringue is the plasma going in circles. Any meringue caught between the two beaters gets accelerated and shot out at high speed--as you've probably discovered when you had to clean meringue off the kitchen wall. When this happens to plasma, we call that high speed spot "the X-line", and the process of speeding up the plasma is called "reconnection".

Now it turns out that the X-line is really thin, and impossible to see from the ground. So we have to get lucky on a NASA safari to see this wild thing. Every NASA chief safari scientist claims to have spotted this elusive prey, and everyone else has been skeptical. So after 50 years, NASA is sending up a special 4-satellite mission in 2017 specifically to encircle this X-line and prove without a doubt that it exists. And we caught it in our bottle almost without trying.

Saturn's Ring

This one came about because we were convinced that strong electric fields were present in the previous two or so pictures. But its hard to see an electric field, and my attempt to measure it with an Etch-a-Sketch was only a partial success. It took about 20 minutes to move the probe (attached to the writing tip of the etch-a-sketch with the front glass removed) but after 7 minutes the magnet overheated and ruined the NIB material. I didn't know it before, but 80C is about as hot as any of those magnets can survive.) I tried to plot up the 3 or 4 points we got, but it wasn't too convincing.

So I came up with another way to measure the electric field--we would sprinkle face powder in the chamber, and the electrified powder would be levitated by the electric field. "It'll never work," I was told, "everybody knows that magnets aren't powerful enough to trap dust." But we ran into a bigger problem--we couldn't see the powder. We shone a laser pointer into the chamber and after many tries, saw a dot. We then vibrated a mirror to make a laser "sheet", and sure enough the dust ring could be seen. To get a photograph, however, we had to nearly point the camera at the laser (forward scattered light is brightest).

In this photograph you see a radiation-belt plasma with mounds of face powder below sitting on an aluminum pie pan. A high voltage spark is charging and throwing face powder upwards. Some of the charged powder remains in a ring around the radiation belt, where a green laser is scanning to illuminate it. My greatest surprise wasn't that it was levitated, but that it was such a narrow ring that looked just like Saturn.

The science editor at Nature was not impressed. "Everybody knows that electric fields have nothing to do with Saturn's rings" he told me. So I never submitted the picture to a journal for publication. Years later, I learned that scientists in the "dust lab" at the University of Colorado had pinned it to their wall.

They spent the next 5 years writing computer models to explain that picture. And we found it in a jar, because we were too dumb to know it couldn't be done.

Milky Way's "Fermi-bubble"

There is another way to make a cusp that doesn't involve two dipoles, and that is to have a funny shaped magnet that is made up of many little dipoles not exactly aligned. Then one piece of that big magnet is like a little dipole, and it can oppose a distant piece of the same extended magnet. A horshoe magnet is an example of that sort of "one-piece" quadrupole. But there is a more important quadrupole that is far more common, in fact so common that there's at least one in your car and probably a half-dozen or more in your house. They are called "solenoids".

A solenoid is like that old gradeschool experiment where you wrapped a wire around a nail five or ten times and hooked it up to a battery to make your first electromagnet. Then if you pulled the nail out, you still had an electromagnet, but with a hollow core, and that is a "solenoid". You have one in your car to engage the starter motor with your engine, and when you turn the ignition the "nail" gets pulled into the solenoid and the gears of the starter motor mesh with the flywheel to start the engine.

But if you got out a magnifying glass, and looked at the wires in the solenoid to see the individual wires, you would see that the magnetic field of one wire, cancels the magnetic field of the neighboring wire, so that a quadrupole null appears right on the "skin" of that wrapped wire. Now if we replace the wires in a solenoid with NIB magnet, sort of like a tin can with the top and bottom removed, then we would have a "magnetic solenoid", where the quadrupole cusp appears along the centerline of the system, but just outside the (missing) top and bottom. If I could draw it, it would have the shape of a Hershey's kiss.

And that is what you are seeing in that picture. The ring of NIB magnet has a "radiation belt" of trapped plasma at the equator, since from far away, the plasma doesn't know the ring is hollow--it looks just like the solid magnet in the first picture at the top. But up close, there's this quadrupole cusp that traps so much glowing plasma, that in this 8-bit photo it saturates the color to white. That's a lot of trapped plasma!

But something else is going on too. Look above the "Hershey's kiss" of trapped cusp plasma, and what do you see? A thin wisp or bubble of plasma. Why is it so thin compared to the bright cusp below it? Because it isn't trapped plasma, its a thin stream of plasma just like the meringue caught between the beaters. It's an accelerated "X-line", which under other circumstances, can form a jet, but in this case it didn't get enough speed to overcome the magnetic field on the opposite side of the cusp, so it just formed a bubble.

Bubbles are significant too. And now here's the rest of the story.

This is a picture taken with the Fermi gamma-ray telescope in orbit around the earth. It is the entire night sky projected onto a big oval, a "Mollweide projection". It's been processed to only include gamma-rays coming from hot electrons (the kind that make up plasmas). What you see are two big bubbles coming out of the center of the Milky Way galaxy, as if all those stars spinning around the galaxy have formed a giant solenoidal magnet. They were so unexpected, they were called "Fermi-bubbles" in honor of the experiment that found them.

And once again, what was just discovered 5 years ago--with a detector launched on our biggest rocket--is larger than our galaxy, but can fit in our tiny bell jar.

Astrophysical Jets

But wait a minute, I didn't call this "the galaxy in a bottle", I said "the universe in a bottle"; what could be bigger than the Milky Way? How about a giant galaxy, with two jets coming out of it megaparsecs long? They're called Active Galactic Nebulae (AGN) and they are up there with galaxy clusters as some of the biggest objects in the universe.

<-------------- 190 Mega-parsecs ------------->
Courtesy NRAO/AUI.

Unlike our picture of a Fermi-bubble in a bottle, these AGN jets are two-sided. For many years, I could never understand why our magnetic solenoid was one-sided. We turned it on and off several times, and perhaps every fifth or tenth time, it formed a 2-sided jet, but most of the time it was one-sided. At the time, I had thought I had made an astrophysical jet, but if astrophysical jets were one-sided, they'd be rocketing all over creation. Now I realize that one crucial feature kept our magnetic solenoid from being an AGN--it had no jet, just a bubble.

How do I make it into a jet? I wondered. By studying pictures of astrophysical jets, I began to see the difference. Where the magnetic solenoid had a "radiation belt" plasma, the AGN had an accreting disk; The magnetic solenoid was making the plasma on the outside, but the AGN was being squeezed by the outside plasma; The magnetic solenoid was pushing out, the AGN was squeezing in.

And that led to the next experiment.

We took the same cylindrical magnet from the very first experiment, and we made an "accretion disk" around it, which we constructed out of braided copper wire with brass safety pins stuck through it. (We tried straight pins, but the magnet pulled them right in.) The sharp points on the brass pins are really good at making plasma because like lightning rods, their electric field is really strong. Now when we put -400VDC on the brass pins, we get lots and lots of plasma in a ring around the magnet. This simulates the AGN accretion disk squeezing the magnet.

So we hooked up two power supplies, one for the magnet (to make a radiation belt), and one for the brass pins (to make an accretion disk). Then we photographed the plasma over the top of the magnet (where the magnetic solenoid formed a cusp). By varying the voltages, we could vary the plasma glow, and we wanted to see if we could make a jet. No such luck.

Next we put a nail on top of the magnet. Now the nail is made of iron, and iron sorta slurps up the magnetic field lines, so it is as if we had replaced our cylindrical magnet with a tall pointy one. Now when we varied the voltages we saw a big change in the shape of the plasma. Okay, it wasn't as dramatic as I had hoped, becase just when it started looking dramatic, all our wires start to spark and jump around because they didn't have enough insulation on them. So we had to make do with less-than-spectacular photos, but that didn't stop us from plotting the extension of the plasma as a function of the voltage.

Yeah, this looks like just another plot, and not too exciting. But what made me sit up and notice was the R2=0.89 correlation coefficient, which MS Excel kindly provides for their "least squares" fit. If you were making predictions on the stock market and got a 0.89 correlation, you would be a millionaire. That's an absolutely awesome correlation. The only bigger correlations are things like death and taxes.

So what does it mean? It means that if we had the ability to put a whole lot more voltage onto our magnet, we'd be shooting plasma out the ends just exactly like an AGN.

And you know what? Astrophysicists have known about AGN's for 50 years and they still don't know how to make them. Even with a computer. And supermassive black holes. And with neutrinos and superconductivity and dark matter to boot. But we did it, with a nail and safety pins, in a jar.

That's why I called this "the Universe in a Bottle".

Thanks to Eric Reynolds, Sudhir Pandkar, Scott Spurrier, Geoff, Chris Best, and ??, all the students who helped make these marvellous pictures. And thanks to Sarah, Keziah, and Malkah who made "the jar" their Science Fair project.
Last updated: 28-Jan-2016