A Tale of Two (Neutron) Stars

At the end of last week, one of our colleagues here in Oxford, Professor Philipp Podsiadlowski, told us that he and some others had discovered something really interesting about neutron stars: how fast they are spinning could be telling us how they were born. This discovery has already hit the blogosphere in quite a big way (here and here and probably elsewhere too). So what’s it all about? Why is it so important? I spoke to Philipp (who is one of three authors of a Nature paper about the discovery) on Monday; here’s what I found out.

A brief history: on the shoulders of giants

For the last 20 years, we’ve known that a lot of neutron stars (abbreviated to NS from now on) have quite high velocities compared to the other stars in our Galaxy. An NS is one of the possible end products when a massive star explodes in a supernova at the end of its life (the other possibility is a black hole). Furthermore, when we find an NS in a binary system, we find that the orbital period (the time taken for one lap) is proportional to the eccentricity (how squashed the orbit is: rugby balls have bigger eccentricities than footballs). For an NS in a binary orbit, the bigger the orbital period, the bigger the eccentricity.

We think we understand this: it all comes down to something called a kick. When a massive star explodes in a supernova, most of the energy is expelled in particles called neutrinos (you may have heard about these little tikes seeming to disrespect the laws of Physics recently, that’s a whole different story). In fact, it’s the neutrinos that make the supernova explode rather than implode: a small fraction of the neutrino energy is absorbed by the other (collapsing, very dense) outer material, which blows it up and out. But if 1-2 % of the neutrino energy is captured more on one side than the other, then the conservation of momentum (Newton’s idea that “action and reaction are equal and opposite”) dictates that the lopsided explosion will launch the remnant neutron star off in the other direction. This is the kick which can cause neutron stars to have large velocities. If the supernova is in a binary system, the kick will likely increase the size of the binary orbit, along with the orbital period and eccentricity. However, not all neutron stars have such high velocities: some have low velocities. What could be going on?

Two formation channels: the theory

This puzzle led some astrophysicists to propose that two different types of supernovae could provide two different “formation channels” for NS’s: the electron-capture mechanism, versus the iron core-collapse mechanism. These two take some serious explaining, so here goes.

Iron core-collapse: this occurs when a massive star burning hydrogen and helium to iron (via nuclear fusion) makes enough iron in its central core to approach the Chandrasehkar limit (that’s when the mass of the iron core reaches 1.4 times the mass of the Sun). When this happens, the iron core collapses uncontrollably: the iron gets so hot and compressed that it breaks up into protons and neutrons, releasing lots of energy (in neutrinos). The neutrons and protons end up in the NS, while the neutrinos fly away and impart a small fraction of their energy into the outer layers of the star, causing it to explode as a supernova..

Electron capture: this is similar to the iron core-collapse case, but this time the (lower mass) star has produced a core made of oxygen, neon and magnesium (or O, Ne & Mg using chemical symbols). This ONeMg core is supported, not by the energy of nuclear fusion, but by a quantum mechanical effect of electrons, called degeneracy pressure (it's very difficult to squash two electrons into the same quantum state). However, when this core reaches the mass of 1.37 times the mass of the Sun, the magnesium and/or Neon start to eat all the electrons! Can you guess what happens next? With no electrons, there’s no pressure support, and the core collapses into a neutron star (with mass around 1.25 times the mass of the Sun), releasing a huge amount of energy in neutrinos which again drive the outer layers to explode in a supernova.

There is one important difference between these two proposed supernova mechanisms: the iron core collapse takes longer to explode because the neutrinos have more work to do (the iron core of the star is more massive). This is important because during the collapse of either star, something called an accretion shock instability begins. Remember how if 1-2 % of the neutrino energy is captured more on one side than the other, the NS gets a kick? This is that same process. In simple terms, the core begins to wobble, and this wobble gets bigger and bigger with time. If the collapse takes a long time, the wobble (and resulting kick) gets very big. This is why NS’s from iron core collapse supernovae should have bigger velocities than NSs from electron-capture SN: they were wobbling for longer during the core collapse!

The Latest: new observations

So, how to test this theory? Are there really two types of supernova, associated with either strongly or weakly kicked neutron stars? What Philipp and his team did was look for another way of separating the two populations - and then to see whether it fit the two supernova picture.

Some NSs live in binary systems, and some of them orbit a special type of star, called a Be star (if you went star gazing with Monday’s skymap, you might have seen one of these already). Be stars expel a lot of gas (they’re young, big and boisterous) and when the NS comes close enough during its orbit, it sucks up some of that gas and emits X-rays in the process. Furthermore, the X-ray emission is pulsed with the spin period of the NS. So in Be/X-ray NS’s, we can measure how fast the NS is spinning quite easily - far more easily than we can measure the orbital period or eccentricity of the binary. The latest work identified that the spin periods of the Be/X-ray NS’s are proportional to the orbital periods of the binaries they live in, and that both the spin and orbital periods seem to be clumped into two groups. There’s a fast spinning, quick orbiting group with low eccentricities as well as a slow spinning, slow orbiting group with high eccentricities. This is the first time such a clear division has been seen in the properties of neutron stars, and it matches the predictions of the two formation channel theory above very well!

What’s even better is that the division is most easily seen in the spin period, which is easy to measure for Be/X-ray binaries. One little point to note though is that these spin periods are not caused directly by the SN explosion, unlike the orbital period or eccentricity. The spins evolve over time to a steady equilibrium state, so it’s likely that the two formation channels create NS’s with different properties (such as magnetic field), which leads to two different equilibrium spin states.

What to do next?

At the moment the data is a little sparse and we need more to really set the result in concrete, but it’s already looking very convincing. In particular we need to measure the orbital periods and eccentricities of more Be/X-ray NS’s: remember this is difficult, so not many have been done (astronomers tend to do the easy things first - can you blame them?). But if the two groups of neutron stars really do come from two different formation channels, then the theory makes some predictions:

1) All the low spin period NSs (whether in binaries or not) should have low velocities: the electron-capture mechanism does not impart large kicks.

2) The two NS groups should have different masses: electron-capture NSs should have masses around 1.25 times the mass of the Sun, while the iron core-collapse NSs should have masses more like 1.4 times the mass of the Sun.

If you’re out star gazing, have a glance at these...

While I was talking to Philipp, he also told me a few other fascinating facts. Firstly, globular clusters (small dense clusters of stars, some of which can be seen with a small telescope) have lots of NS’s compared to the average in our Galaxy. Most of these NS's probably formed via the electron-capture channel in dynamical interactions, otherwise the kicks imparted from the iron core-collapse mechanism would have kicked them out of the cluster! Secondly, the Crab nebula (sometimes called M1, shown in the image) is consistent with being an electron-capture NS. This NS was formed from a supernova explosion in the year 1054 and was noted by Chinese, Japanese and Arabic astronomers at the time. It would have been so bright then, that you could have seen it in the day and it would have lit up the night like a second moon! Wow. So don’t go thinking all this stuff has no relevance to you. We might see an electron-capture supernova in the sky tomorrow! Then again, I might win the lottery :-)

Image credit: NASA/CXC/SAO/F. Seward et al.