by Lucie Green
Solar flares themselves were suggested as a possible energy source for CMEs because it was noticed that a lot of the time when a CME occurs, there is also a solar flare underneath. It was thought that the solar flares could somehow be blasting away the plasma above them. Solar flares release a vast amount of energy within minutes, which goes into heating the plasma in the corona to tens of millions of degrees. And the thermal energy, because of the plasma temperature, might provide the energy for the ejection and a large pressure that could expel the material.
But it was soon realized that there simply isn’t enough thermal energy available in solar flares to do this. Moreover, flares and coronal mass ejections do not always occur together. Many flares have no ejections, and many ejections have no flares. We needed a mechanism to drive a coronal mass ejection that worked without an assisting solar flare.
One by one, possible energy sources were eliminated. Until the only one left was, once again, energy stored in the magnetic field.
I’m not sure which came first: my obsession with magnetic fields or my love of the mysteries behind CMEs, but they soon became intertwined for me. Work on solar flares had shown how energy could be released from magnetic fields through magnetic reconnection, as the magnetic field rearranged into a lower-energy state. A small region of reconnection could have large consequences and cause electrons and protons to rush down the magnetic field to produce all the aspects of solar flares we observe. Could magnetic energy be used to liberate something as massive as a CME?
The breakthrough is to realize that the mass in ‘coronal mass ejection’ is perhaps something of a distraction. Let’s not focus on the mass for the moment. Actually a CME is an eruption of magnetic field escaping from the Sun that races into the Solar System. The mass just happens to be there for the ride. It is like looking at a tree moving about on a windy day and concluding it is pushing the air around. It’s actually the air that is moving and the tree is merely being swept along with it.
CMEs are first and foremost the Sun’s magnetic field evolving. But since the magnetic field and the plasma are frozen together, the plasma must also be swept upward. One of my colleagues, Tom Berger, is on a semi-serious campaign to rename CMEs as ‘coronal magnetic eruptions’. I think he also likes the fact that Hale, discoverer of the Sun’s magnetic field, used to refer to flares as ‘magnetic eruptions’. Tom is perpetuating the phrase used by the first solar astrophysicist, albeit with a modern and accurate meaning. Which is not to downplay the significance of the plasma. But the hunt was now on for what could cause the magnetic field to suddenly erupt.
SECOND TIME AROUND
The rapid changes in the magnetic field of the solar atmosphere, which lead to a vast bubble of it being ejected as a CME, reveal that there are times when it is highly dynamic. It has get-up-and-go on the most colossal scale and is in stark contrast to any magnetic field that we might interact with here on Earth. And this dynamic side needed to be explained. The solution came down to the very same magnetic fields we originally used to explain sunspots: ropes of magnetic field. In our explanation of sunspots, the flux ropes become buoyant and rise from the tachocline up to and through the photosphere. But to use these flux ropes to understand CMEs, we actually need to fix a problem I sneakily skipped over before.
Flux ropes do indeed form down at the tachocline because of the different rotational speeds within the sun stretching, amplifying and twisting the magnetic field. The plasma within them then becomes buoyant and they float up to the photosphere. At this point I said that they protrude up into the solar atmosphere and that where the magnetic field comes out of (and goes back into) the photosphere is where we get sunspots. All true, but the hard part is actually getting the flux ropes to push through the photosphere. The situation isn’t as simple as I made it seem.
The less-dense plasma inside a flux rope makes it buoyant enough to rise up to the photosphere, but not buoyant enough to pass through it: the surrounding plasma changes too dramatically. Compared to the plasma inside the Sun, the plasma trapped inside a flux rope is light, but compared to the plasma in the atmosphere it’s rather heavy. If you release a piece of wood under water it will float up to the surface because it is less dense than water. But the wood will not rise out of the water and float up through the air. It will bob around on the surface. This is what flux ropes do just under the photosphere.
This is all happening beneath the photosphere – out of our direct view. But this doesn’t stop us believing that we have good reason to think this is what is happening – although it seems like a paradox: the photosphere stops the magnetic field, yet we see it breaking through to form sunspots. If you look at sunspot pairs as they form though, you see tiny pieces of magnetic field breaking through that gradually over time build up to form the spots. But it isn’t the whole flux rope that emerges. Much like a floating piece of wood ends up half under and half above the water’s surface, the flux ropes will end up partly above the surface and partly below. It is these half flux ropes that form the magnetic arches and arcades seen above the photosphere. But some (it’s not known how much!) of the original flux rope always stays below the surface. (See plate 17.)
The reason why this partial emergence of the flux rope is disappointing for understanding CMEs is that if we could have a complete flux rope entirely above the photosphere, it could explain how CMEs are produced. Magnetic fields can hold energy if they are twisted, and so if there were a twisted flux rope, composed of a bundle of helical field lines, that had been bent into an arch, it could be bursting at the seams trying to spring back and straighten out the magnetic field. Flux ropes are also cohesive structures that can easily trap plasma. And so they would explain why CMEs are such a massive movement of plasma and magnetic field simultaneously: all that plasma could be trapped in the same flux rope.
We just need some way for flux ropes to form above the photosphere and be held there until suddenly released.
My interest was in working out whether magnetic reconnection in the photosphere could form twisted flux ropes that are tethered down by other magnetic arches. In short: a second-generation flux rope can form in the magnetic arches left over from the first-generation flux rope half poking through the solar surface. This idea was already around but needed testing against observations.
All that is left of the original flux rope is a series of arches of magnetic field that come out of the photosphere and curve around to come back down. But if two of those arches meet at their feet in (or close to) the photosphere, they can reconnect and form a helical field line. A small loop is also produced underneath the helical one, and this submerges below the photosphere. If a series of arches continue to meet and reconnect, a new flux rope will form as more arches reconnect into it. Importantly, there will be plenty of magnetic arches still reaching over this flux rope, holding it in place. At least, for a while.
Eventually, the energy in the twisted flux rope becomes too much and the magnetic arches over the top of it can no longer hold it in place. The upward force on the flux rope is caused by the magnetic field being more concentrated low down in the atmosphere than it is higher up, creating a steep gradient in the magnetic pressure. The magnetic arches succumb to the upward force of the flux rope and are dragged into the Solar System. The flux rope breaks away from the Sun, resulting in the sub-escape-velocity speeds we observe – although some flux ropes involve so much energy that when they escape they move at phenomenal speeds.
12.2 The creation of a flux rope from an arcade of magnetic field lines rooted in a moving photosphere. Magnetic reconnection happens between panels (c) and (d) between arch foot-p
oints B and C. Magnetic reconnection happens again between panels (e) and (f) and this time involves foot-points F and G (adapted from van Ballegooijen and Martens (1989)).
This is all great in theory. But how do we know that flux ropes are definitely responsible for CMEs? We’d need to actually observe them in the corona. Frustratingly, it is very hard to see the magnetic field throughout the corona. The million-degree plasma is so hot it smears out any signature of the magnetic field that might be sent to us in the radiation. Magnetic fields are more readily measured in the relatively cool plasma of the photosphere.
So we need to use the age-old scientific test: see if our theory makes any predictions and then test those predictions. We need to find something uniquely ‘flux-ropey’ and see if there are signs of that in CMEs. Thankfully, the helical field lines in a flux rope gives us just such a property. Flux ropes can be twisted in two different directions: clockwise and anti-clockwise, which we actually call right-and left-handed directions of twist. And flux ropes with different directions of twist can behave slightly differently when they erupt.
Previously, solar scientists in America had noticed that the plasma in the corona would sometimes form an ‘S’ shape and that this indicated it was highly likely to erupt and produce a coronal mass ejection. And they had the idea that these ‘S’ shapes – which have become known as ‘sigmoids’ – are magnetic flux ropes. (See plate 18.) And, if they are flux ropes, the shape of a sigmoid would indicate the direction of the flux rope’s twist. A normal ‘S’-shaped flux rope would have a right-handed twist, whereas a mirror-image ‘S’ (-shaped) would be left-handed. These details are needed to test the idea that sigmoids are the observational signature of flux ropes. And it is an exciting idea because if we can find flux ropes in the solar atmosphere, we might be able to solve the two main mysteries of CMEs at once: where does the energy come from and why does the magnetic field become so mobile?
Working with my colleagues, including Bernhard Kliem, who had worked on the theory of how flux ropes erupt, we began testing theoretical ideas. Firstly, that sigmoids form on the Sun as the result of magnetic reconnection between the arches of a partially emerged flux rope. Here we found many cases where the observations very nicely matched the theory – a good indication that these sigmoids are actually flux ropes! But there’s more. The theoretical work made predictions about how flux ropes should behave. One example is that a flux rope which has a left-handed magnetic field twist should behave in a slightly different way when it erupts to a magnetic flux rope that is twisted in a right-handed sense. The curvature of the field lines of the flux rope means that a left-handed flux rope should, as it erupts, rotate in a counter-clockwise direction. A right-handed rope will rotate clockwise – albeit often only by a small amount.
So an erupting flux rope formed in an -shaped sigmoid should rotate in a different direction to one formed in the normal S-shaped ones. We scoured observations and did a survey of which way the CMEs rotated when they erupted. Sure enough, the observations matched what the theory had predicted: the direction of twist in a sigmoid was related to how it rotated, supporting the interpretation that they are magnetic flux ropes.
There are still plenty of mysteries about CMEs and flux ropes for us to work on though. One is the exact role of magnetic reconnection. I said that in many CMEs flares occur as the eruption takes place. And flares are the result of magnetic reconnection. What’s not clear is the order of these two events – our observations are not yet good enough to see whether the magnetic reconnection, and solar flare, actually begins before or after the CME erupts. It could be that the CME occurs first, causing the reconnection to take place. As always, there is much we still need to learn about the Sun.
THE FINAL TWIST
Coronal mass ejections go from being a curious feature of the corona to a vital part of the Sun’s magnetic field evolution when you look at an obscure aspect of magnetic fields: helicity.
Helicity is a curious property of magnetism that isn’t often talked about. It is a general measure of how distorted a magnetic field is. This distortion is ultimately responsible for the magnetic energy stored in the fields, but magnetic helicity is useful in its own right because it provides a way to measure exactly how distorted the magnetic fields are: how twisted, braided or interlinked. The plasma flows in the photosphere and below can lead to a high degree of distortion in the magnetic field of the Sun, and helicity provides a way to track how magnetic distortion moves through the Sun from the tachocline to the corona.
The concept of helicity actually goes right back to Karl Friedrich Gauss, the early-nineteenth-century mathematician, astronomer and physicist who contributed greatly to the theory of electromagnetism – the strength of a magnetic field is measured in units of ‘gauss’ in his honour. But the reason why magnetic helicity never became widely known about is that for a long time it remained an abstract concept. When I started work, magnetic helicity was on the periphery of most solar physicists’ attention. Today, it’s a fairly commonly used phrase.
12.3 A coronal mass ejection imaged by SOHO showing a helical structure in the plasma. This shape is probably caused by a twisted magnetic field known as a flux rope (Courtesy of ESA/NASA and the LASCO consortium).
Helicity is incredibly important for solar physics for two reasons. Firstly, because the tangles in the Sun’s magnetic field are so important, and, secondly, because magnetic helicity is such a tenacious property. Once magnetic field becomes tangled, there is no easy way for that magnetic helicity to be lost. We say that magnetic helicity is ‘conserved’. Even during something as dramatic as a solar flare, there is the same amount of magnetic helicity after all that magnetic reconnection as there was before.
It can take a process as extreme as the rotation inside the Sun to produce a twisted and distorted magnetic field. One set of ideas suggests that the Sun’s tachocline is a magnetic helicity generator, pumping out twisted magnetic field. But magnetic helicity can also be injected at other stages of a magnetic field’s lifetime – by plasma flows in the convection zone, for example. Observations suggest that by the time the magnetic field emerges into the corona, it is twisted and distorted. And remember that this magnetic field keeps on emerging over the solar cycle.
My own interest in the role of magnetic helicity in CMEs came about when I read a paper written by a mathematician and solar physicist in the US called Boon Chye Low. ‘BC’, as he is known, had studied mathematics at the University of London but by the time of this paper had spent many years working at the High Altitude Observatory in Boulder, Colorado. The paper looked at the very fundamentals of coronal mass ejections. He thought about why the Sun might be producing such impressive ejections, which led on to how they might be happening.
BC wrote that coronal mass ejections were the way that the Sun could shed magnetic helicity and prevent it from being endlessly accumulated in the corona over each solar cycle because of the ongoing emergence of magnetic flux into the atmosphere. Coronal mass ejections appeared to be a valve to remove magnetic energy that was stored in the twisted and distorted magnetic fields of the corona and couldn’t be released any other way. It seems that coronal mass ejections are an essential part of the magnetic cycle of the Sun. In fact, BC wasn’t the only person thinking along these lines. David Rust, who had put forward the idea that sigmoids were flux ropes, also suggested that CMEs were necessary for shedding magnetic helicity from the corona.
But just how much magnetic helicity is shed? Answering this was a major challenge. But in the early 2000s magnetic helicity went from being an abstract theoretical concept to being a quantity tha
t could be worked out using observations. A small number of pioneering solar physicists developed ways to use observations of the Sun to follow magnetic helicity and I had the honour of working with two of them.
Pascal Démoulin works at the Paris Observatory, as he did when I first met him in 1999. As I know he will read this book, I won’t say too much about my first memories of meeting him, but the mention of magnetic field caused his eyes to light up and sparkle through an otherwise dominating beard. Pascal specializes in using mathematics and computers to investigate the shapes solar magnetic fields could form. He had already modelled the magnetic configurations likely to lead to flares and was about to turn his mind to CMEs.
As we saw earlier, using the fact that magnetic fields leave their signature in the light coming from the Sun means that we can make a magnetic map of the solar surface. But this is only a thin slice through the Sun’s magnetic field at the photosphere level; we cannot directly measure the magnetic field lines as they stretch up into the corona. Which is why Pascal’s models are so important: he can take the magnetic data from the photosphere and mathematically simulate what it is likely to look like above that. Observations of plasma in the corona which trace out the magnetic structures can be used to check that his calculations are accurate.
Meanwhile, at UCL we had one of the world experts in helicity: Mitch Berger. The work of Mitch and Pascal meant we were able to calculate how much magnetic helicity there was in certain parts of the coronal magnetic field. Then we could take observations of CMEs and calculate how much magnetic helicity they were carrying away from the Sun. Which we could then match up to what we had seen on the Sun from the area they originated from.