Novel interpretation of the latest AMS-02 cosmic-ray electron spectrum

The latest AMS-02 data on cosmic-ray electrons show a break in the energy spectrum around 40 GeV, with a change in the slope of about 0.1. We perform a combined fit to the newest AMS-02 positron and electron flux data above 10 GeV using a semianalytical diffusion model where sources include production of pairs from pulsar wind nebulae (PWNe), electrons from supernova remnants (SNRs), and both species from spallation of hadronic cosmic rays with interstellar medium atoms. We demonstrate that within our setup the change in the slope in the AMS-02 electron data is well explained by the interplay between the flux contributions from SNRs and PWNe. In fact, the relative contribution to the data of these two populations changes by a factor of about 13 from 10 to 1000 GeV. The PWN contribution has a significance of at least $4\sigma$, depending on the model used for the propagation, interstellar radiation field, and energy losses. We check the stability of this result against low-energy effects by numerically solving the transport equation. as well as adding possible breaks in the injection spectrum of SNRs. The effect of the energy losses alone, when the inverse Compton scattering is properly computed within a fully numerical treatment of the Klein-Nishina cross section, cannot explain the break in the $e^{-}$ flux data, as recently proposed in the literature.

Figure 1

Left: comparison between the spectrum of the ISRF density $n(\epsilon )$ published in Ref. [11] (blue dot-dashed line) with the blackbody approximation introduced in this paper (dotted orange line). Right: comparison among the ISRF models Porter2006 [15] (solid black line), Vernetto2016 [11] (blue dot-dashed line), Delahaye2010 [16] (dashed green line), and Evoli10/2020 [14] (dotted red line).

Figure 2

Energy loss rate $b(E)=dE/dt$ (multiplied by $E^2$) for ICS off the ISRF photons composed of CMB (green lines), dust emission (red lines), and starlight (blue lines) for $e^{\pm }$ energy $E$. The total rate is shown with black lines. We report three cases: blackbody approximations of the ISRF and approximated Klein-Nishina calculation as in Ref. [13] (dotted lines), blackbody approximations of the ISRF and full numerical Klein-Nishina calculation (solid lines), and the Vernetto2016 ISRF model and full numerical Klein-Nishina calculation (no approximations, dashed line).

Figure 3

Flux of $e^{-}$ from a smooth distribution of SNRs calculated for ${\gamma }_{\mathrm{SNR}}=2.55$. We show the same cases for the ISRF and tKlein-Nishina energy loss rate as the ones reported in Fig. 2, in order to demonstrate the effect of the approximated calculation of the ICS energy losses published in Ref. [13] and implemented in Ref. [10] on the $e^{-}$ flux.

Figure 4

Left: result for the combined fit to $e^{-}$ and $e^{+}$ AMS-02 data (black and grey data points). We show the secondary production of $e^{+}$ (dashed green line) and $e^{-}$ (dotted green line), $e^{\pm }$ from PWNe (solid red line), and $e^{-}$ from SNRs (dot-dashed blue line). Right: same as the left panel but zooming in on the $e^{-}$ sector.

Figure 5

Flux of $e^{-}$ from SNRs (blue dot-dashed line), PWNe (red solid line), and secondary production (green dotted line) as derived from a combined fit to the $e^{\pm }$ AMS-02 data. We also show the total contribution (black dashed line) and the AMS-02 data (black data points). Each plot refers to one of the first six cases reported in Table 2.

Figure 6

The left (right) panel shows the comparison (ratio) between the numerical calculation of $\mathcal{J}(\mathrm{\Gamma })$ and the approximation made in Ref. [13] as a function of $\mathrm{\Gamma }$.