Pulsars are neutron stars that emit amazingly regular, short bursts of radio waves, so regular that they were originally thought to be signals from little green men! Though their discovery over 40 years ago was very widely reported and resulted in a Nobel Prize, the reasons how and why they send these bursts has remained a mystery; to quote Jean Eilek of NRAO, "we know why they pulse, but why do they shine?"

However, in papers presented this week to the American Astronomical Society, Andrea Schmidt and John Singleton of Los Alamos National Laboratory provide detailed analyses of several pieces of observational data that suggest that pulsars emit the electromagnetic equivalent of the well-known "sonic boom" from accelerating supersonic aircraft. Just as the "boom" can be very loud a long way from the aircraft, the analogous signals from the pulsar remain intense over very long distances.

Schmidt and Singleton's presentations provide strong support for a pulsar emission mechanism (the superluminal model) due to circulating polarization currents that travel faster than the speed of light. These superluminal polarization currents are disturbances in the pulsar's plasma atmosphere in which oppositely-charged particles are displaced by small amounts in opposite directions; they are induced by the neutron star's rotating magnetic field.

Despite the large speed of the polarization current itself, the small displacements of the charged particles that make it up means that their velocities remain slower than light, so that Einstein's theory of Special Relativity is not violated. No laws of physics are broken in this model!

Back in the 1980s, Nobel laureate Vitaly Ginzburg and colleagues showed that such faster-than-light polarization currents will act as sources of electromagnetic radiation. Since then, the theory has been developed by Houshang Ardavan of Cambridge University, UK, and several ground-based demonstrations of the principle have been carried out in the United Kingdom, Russia and the USA.

Thus far, polarization currents traveling at up to six times the speed of light (i.e. 1.8 million km per second) have been demonstrated to emit tightly-focused bursts of radiation by the ground-based experiments. Note that, though the source of radiation exceeds the speed of light, the emitted radiation travels at the normal light speed once it leaves the source.

In the superluminal model of pulsars described by Schmidt and Singleton, the polarization current moves in a circular orbit, and its emitted radiation is therefore in some ways analogous to that of the electron synchrotron facilities used to produce radiation from the far-infrared to X-ray for experiments in biology and other subjects. In other words, the pulsar is a very broadband source of radiation. However, the fact that the source moves faster than the speed of light results in a flux that oscillates as a function of frequency.

Schmidt's presentation, produced in collaboration with Houshang Ardavan, shows that these predicted oscillations are seen in GHz data from the Crab pulsar published by Tim Hankins and colleagues at NRAO. Schmidt also showed fits of the superluminal model to data from the Crab and eight other pulsars, spanning electromagnetic frequencies from the radio to X-rays.

In each case, the superluminal model accounted for the entire data set over 16 orders of magnitude of frequency with essentially only two adjustable parameters. In contrast to previous attempts, where several disparate models have been used to fit small frequency ranges of pulsar spectra, Schmidt showed that a single emission process can account for the whole of the pulsar's spectrum.

Another prediction of the superluminal model for pulsars is that there should be a component of the pulsar's flux that decays as 1/distance, rather than as the conventional inverse-square law. The effect is in fact a general property of sources that both exceed the speed of their emitted waves and accelerate, and has been known in the field of acoustics since the advent of supersonic aircraft; it results from focusing of the emitted waves in the time domain.

In pulsars, the acceleration is centripetal, due to the fact that the superluminal polarization current rotates with the neutron star's magnetic field. Singleton's presentation suggests that this non-spherically-decaying radiation is detected in pulsar observational data.

Along with Pinaki Sengupta (Nanyang Technological University, Singapore), Singleton has analyzed flux and dispersion data from 971 pulsars in the Parkes Multibeam Survey. The analysis, which is derived from a widely-accepted approach originally used by George Efstathiou (Cambridge University) to study the red-shifts of very distant objects, shows that pulsars possess a flux that falls off as the inverse of the distance, rather than the inverse of the distance squared.

This dependence, which is unique to pulsars (and the ground-based radio experiments mentioned above) is caused by the electromagnetic equivalent of the "sonic boom". In other words, the pulsar radio pulses are a natural electromagnetic equivalent of a well-known phenomenon in acoustics.

There are several predictions of the superluminal model which should be confirmed by further observations. First, the overall spectrum of pulsars should be generic, due to the superluminal nature of the source; detail differences are due to differences in the pulsar atmospheres.

The model predicts that the entire frequency spectrum should scale as the cube of the pulsar's rotation period, so that millisecond pulsars will emit only in the radio bands, whereas slow pulsars will emit right out to X-rays. Although this does seem to be true for the data in hand so far, the number of broadband pulsar measurements is very restricted at present. Second, other pulsars should show banded emission in the radiofrequency end of the spectrum, similar to those described by Hankins and colleagues in their Crab observations.

This work is supported by the US Department of Energy through the Los Alamos LDRD program. For further technical information, see preprints at arXiv:0912.0350, arXiv:0908.1349 and arXiv:0903.0399