Precision cosmic-ray data challenge a paradigm

A century after Victor Hess discovered cosmic rays, the question of their sources remains unsettled. Leaving aside the rare CRs with energies E above 10¹⁰ GeV, which probably originate in some sort of monstrous extragalactic accelerators, astrophysicists still don’t know with confidence where most of the nuclei that dominate the intragalactic CR flux are accelerated to their more modest energies. Figure 1 shows those fluxes as functions of the kinetic energy K = E – mc ² with which a nucleus of mass m arrives at the top of the atmosphere. About 95% of them are simply protons, and 90% of the rest are α particles—nuclei of helium-4, the second most abundant species in the cosmos.

The most extensively studied and cited model for the creation of intragalactic CRs is diffuse shock acceleration (DSA) in the expanding outer shock fronts of supernova remnants. In DSA, repeated scattering off magnetic turbulence in the shock front can presumably raise a small fraction of nuclei to 10⁶ GeV (see PHYSICS TODAY, January 2005, page 19 ).

The presumption that DSA in supernova shock fronts dominates the CR flux up to 10⁶ GeV has become something of a paradigm. But the direct evidence for the acceleration of nuclei in those shock fronts is meager and equivocal. Until recently, a strong reason for believing in the paradigm has been the apparently universal power-law falloff with increasing energy all the way out to 10⁶ GeV, as seen in figure 1. In that 2007 data compilation, all the flux falloffs above a mass-dependent threshold are well approximated by E ^–γ with the same spectral index γ of about 2.7. And that sort of structureless universality is precisely what the paradigm predicts.

More recently, however, data from balloon-borne experiments have suggested that the CR spectra are neither so structureless nor so universal. ¹ And now the European collaboration that operates the PAMELA orbiting CR spectrometer has reported high-precision hydrogen and helium spectra that confirm and detail the balloon claims. ² The once attractively simple picture has become complicated. The good news, says Piergiorgio Picozza (University of Rome, Tor Vergata), who heads the PAMELA collaboration, “is that the new results provide hard numbers for testing new models.” The data seem to call for a diverse variety of acceleration sites and mechanisms within the Milky Way.

PAMELA

Launched in 2006, PAMELA is a satellite-based instrument dedicated to precision measurement of CR spectra. Its principal component is a magnetic spectrometer with silicon-strip tracking. Most of the earlier data were taken with hadron calorimeters aboard balloons. PAMELA’s significantly higher precision in its measurement range (1 to 10³ GeV) is due to flying above the atmosphere, which avoids obfuscating consequences of CR collisions; to much longer flight duration and therefore better statistics; and to the magnetic spectrometer, which measures momentum p per unit charge directly without the uncertainties attending the calibration of calorimeters.

Momentum per unit charge is, in fact, the relevant variable for studying the acceleration of CR nuclei and their subsequent propagation through the galaxy’s jumble of magnetic fields. That’s because the path of a bare nucleus of atomic number Z in an astrophysical magnetic field B has a curvature radius given by pc/ZeB. Measuring the curvature of an incident nucleus in the spectrometer’s magnetic field, PAMELA determines its “magnetic rigidity” R ≡ pc/Ze. The detector distinguishes α particles from protons at the same R by their greater ionizing power.

In figure 2, the ratio of the H and He fluxes measured by PAMELA is plotted as a function of R. (At energies high enough for the mass to be negligible, R in GV is numerically indistinguishable from E/Z in GeV.) The flux ratio from spectra measured in the same experiment is particularly informative, because many systematic uncertainties in the individual spectra cancel out in the quotient.

If the spectral indices of the H and He spectra in their power-law regimes (which begin at about R = 30 GV) were identical, as predicted for DSA in supernova remnants, the ratio of their fluxes would be constant. Figure 2 shows that they are not. The ratio falls with increasing R, showing that the H spectrum is steeper. Beyond 5 GV that fall is well fitted by its own power law, whose spectral index

Δγ = γ _H – γ _He = 0.101 ± 0.001

confirms that the H and He spectral indices are distinctly different. The latter more closely approximates the “universal” 2.7.

The flux ratio from a single experiment yields other insights not evident in individual spectra separately measured. Below about R = 30 GV, all the CR spectra in figure 1 turn downward from the power-law dependence they exhibit at higher energies. That turnover is attributed to a local “solar modulation” effect: CR nuclei entering the inner solar system have to make their way against a stiff plasma wind emanating from the Sun. The least energetic CRs are decelerated most, and the effect depends on the Sun’s variable activity. In the absence of solar modulation, the DSA model predicts that the spectra would exhibit power-law behavior down to just a few GeV.

That’s hard for an observer inside the solar system to check. In figure 2, however, the power-law fit to the H/He flux ratio is seen to extend all the way down to R = 5 GV, well into the low-rigidity region where spectra are so distorted by the solar wind that they usually reveal little of their true interstellar character. But because PAMELA’s H and He spectra were measured over the same two-year time interval (incidentally, a period of minimal solar activity), the solar wind at any given R would have affected them equally and left their ratio unaffected.

Beyond the paradigm

Figure 2 also shows an attempt to fit the flux ratio with a model put forward by Victor Zatsepin and Natalia Sokolskaya (Moscow State University) in response to indications from early balloon data. ³ It posits additional Milky Way sources of cosmic rays besides isolated supernova remnants: novae and superbubbles. The former are thermonuclear explosions survived by the white dwarf stars that host them, and the latter are bubbles of hot gas, hundreds of light-years across, blown by coherent supernova explosions in clusters of supergiant stars.

In addition to allowing different nuclear species to exhibit different spectral indices, the Zatsepin model accommodates breaks in the spectral index for a single species. Such breaks had been suggested by balloon data, and now they have been pinpointed by PAMELA. Figure 3 shows sharp breaks in both the H and He spectra at about the same rigidity, R ≈ 240 GV. Both spectra curve downward from their lower-energy power-law falloff just before the break and then abruptly “harden” to slower falloff. Such breaks might well be indicating different classes of intragalactic CR sources with different upper limits of acceleration.

It may even be that one should not, in general, think of individual sources for individual intragalactic CRs. Addressing what he characterizes as “the myth” of supernova-remnant predominance, theorist Yousaf Butt (Harvard–Smithsonian Center for Astrophysics) has suggested that the whole galaxy and its extended halo should be considered a “single holistic acceleration site.’’ ⁴ He argues that any given CR nucleus has probably been accelerated and reaccelerated over galactic distances by a variety of sources such as supernova remnants, superbubbles, plasma-reconnection events, and the shock front that marks the outer termination of the galactic plasma wind.