MagnetarsWe now have our second “frontiers” lecture. This time, the subject is magnetars. Theidea is that some neutron stars may have such strong magnetic fields, B > 1014−15G, thata variety of exotic quantum electrodynamic processes, which are otherwise negligible, maybecome dominant. Since these processes cannot be observed in terrestrial lab oratories,magnetars may be our only chance for the forseeable future to test the predictions ofQED in this extreme environment. We’ll start by describing the evidence circa 1995 thatmagnetars exist, then talk about more recent developments, then discuss some of the QEDeffects themselves.Early evidence for magnetarsTo understand the early evidence for magnetars we have to have a brief discussion ab ou tgamma-ray bursts (which we’ll talk about in much more detail at the end of th e course).Gamma-ray bursts are, well, bursts of gamma rays that last, typically, a few seconds. Theywere discovered in the late 1960s by orbiting satellites designed to determine if the USSRwas testing nuclear weapons in space. In the late 1970s and early 1980s, it was discoveredthat some of the sources repeated, although most did not. It now seems that the repeatingsources have a number of differences from classical gamma-ray bursts, so they deserve theirown class designation. The repeating sources (1) repeat (irregularly), (2) have relativelysoft spectra, meaning spectral peaks in the 20-30 keV range instead of 200-300 keV, and(3) have bursts that are typically short, maybe a tenth of a second or less. This class goesby the name “soft gamma-ray repeaters”, or SGRs for short. However, like many anotherclass in astronomy (high-energy astronomy in particular), this is a rather sparse class: onlyfour SGRs are known, with a fifth possible candidate!The fact that SGRs repeat meant that their locations could be identified much moreaccurately than the locations of classical GRBs, which are essentially one and done. Thislocation led to the early conclusion that all of them were associated with supernovaremnants. Ask class: if they were associated with supernova remnants, what kind ofobjects could they be? Neutron stars or black holes. Ask class: if the supernova remnantswere still there, what does this say about the ages of the objects? Supernova remnants arenot visible after ∼ 105yr, so these had to be relatively young objects.Additional analysis indicated that at least a couple of the SGRs were not in the centerof the remnant. Assuming that the birthplace of the SGRs was the center, and using theinferred age of the remnants, this suggested a very high transverse velocity, more than1000 km s−1. Assuming the SGRs were ind eed in the supernova remnants, the persistentand bursting flux (assuming isotropic emission) implied that the persistent luminosity wastypically 1035−36erg s−1and the bursting luminosity was 1039−42erg s−1, depending on theburst. During bursts, the peak emission remains in the 20-30 keV range. Finally, one ofthe SGRs, in the Large Magellanic Cloud, had one burst dramatically different from thetypical burst, on 5 March 1979. For one thing, it was extraordinarily luminous: the peakflux implied an isotropic luminosity of 2 ×1045erg s−1! It was also an extremely long burst,lasting over 200 seconds. In addition, during the long tail clear pulsations with a period of8 seconds were discovered.In 1995, Chris Thompson and Rob Duncan considered this evidence. The following is asimplified version of their arguments and conclusions.First, the regularity of the pulsations in the March 5 light curve indicates that theseobjects are neutron stars, not black holes. As with the argument that pulsars are not blackholes, black holes simply don’t have any structure that a regular pulsation could grab onto.Next, what could be the ultimate power source of the emission? We need to determinewhat energy has to be explained. Ask class: how do we do this? Start with the persistentluminosity. With 1036erg s−1over 20 years, that’s 6 × 1044erg, and if you assume thatthis emission has been typical over the ∼ 103yr age of the supernova remnant, th at’s3 × 1046erg. One must also consider the energy in the March 5 event: there had to beenough energy to power one 1045erg event at that time.Ask class: what energy reservoirs can you think of? Rotation, accretion, and nuclearenergy are all important for one n eutron star or another, so let’s consider them. Whatabout rotation? The total energy of a rotating object is12IΩ2. Assuming that the 8 s periodof the March 5 event was its rotation period (and it is consistent with that interpretation),and using I = 1045cgs, the energy is only about 3 × 1044erg. This is not enough to powerthe March 5 event, and is also not enough to power the subsequent persistent emission.Therefore, argued Thompson and Duncan, r otation can be ruled out.What about accretion? Ask class: given how Bondi-Hoyle accretion works, do theyexpect a high accretion rate from an object moving at 1000 km s−1? No way! The accretionrate goes like v−3, so such a high-velocity object accretes practically nothing from theinterstellar medium (around 106g s−1, in fact, giving a paltry 1026erg s−1maximum). Theonly way that accretion could be a significant source of energy would be if the neutronstar brought along a companion with it. But with such an enormous velocity, presumablyproduced in the supernova explosion that formed the neutron star (pre-SN high-mass stellarsystems are very low velocity), it would be almost impossible to carry along a companion.For this and other reasons, Thompson and Duncan argued that accretion can also be ruledout.What about nuclear en ergy generation? This is a non-starter, for several reasons. Askclass: how can nuclear energy generation be ruled out as the primary source of energyfor SGRs? One p roblem is that there is nowhere that the fuel could come from, if it isn’taccreting actively. Anoth er p roblem is t hat the ultrahigh luminosity of the bursts, muchabove Eddington, does not mesh with a nuclear burst, or accretion for that matter, unlessthe magnetic field is extremely h igh.What about magnetic fields? This was the model p roposed by Thompson and Duncan.Suppose that somehow a few neutron stars have dipolar fields on the order of B = 1015G.If this is approximately the average th roughout the star, then the total magnetic energyis EB= (B2/8π)(4πR3/3) ≈ 1047erg. If this energy is converted into
View Full Document