Evidence of TeV halos around millisecond pulsars

Using data from the HAWC gamma-ray telescope, we have studied a sample of 37 millisecond pulsars (MSPs), selected for their spindown power and proximity. From among these MSP, we have identified four which favor the presence of very high-energy gamma-ray emission at a level of $(2\mathrm{\Delta }\ln \mathcal{L}{)}^{1/2}\ge 2.5$. Adopting a correlation between the spindown power and gamma-ray luminosity of each pulsar, we performed a stacked likelihood analysis of these 37 MSPs, finding that the data supports the conclusion that these sources emit very high-energy gamma-rays at a level of $(2\mathrm{\Delta }\ln \mathcal{L}{)}^{1/2}=4.24$. Among sets of randomly selected sky locations within HAWC’s field-of-view, less than 1% of such realizations yielded such high statistical significance. Our analysis suggests that MSPs produce very high-energy gamma-ray emission with a similar efficiency to that observed from the Geminga TeV-halo, ${\eta }_{\mathrm{MSP}}=(0.39-1.08)\times {\eta }_{\text{Geminga}}$. This conclusion poses a significant challenge for pulsar interpretations of the Galactic Center gamma-ray excess, as it suggests that any population of MSPs potentially capable of producing the GeV excess would also produce TeV-scale emission in excess of that observed by HESS from this region. Future observations by CTA will be able to substantially clarify this situation.

Figure 1

The value of $(\mathrm{TS}{)}^{1/2}$ found using the 3HWC Survey Tool in random sky locations near the faint HAWC source 3HWC $\mathrm{J}1743+149$ ($\mathrm{TS}=25.9$), and the bright HAWC source 3HWC $\mathrm{J}1104+381$ ($\mathrm{TS}=3025.3$). In removing any MSP from our analysis that is located within $0.5°\times [\ln ({\mathrm{TS}}_{3\mathrm{HWC}}){]}^{1/2}$ of any source in the 3HWC catalog (corresponding to the vertical dashed lines), we significantly limit the potential for contamination from this collection of sources.

Figure 2

The gamma-ray fluxes (evaluated at 7 TeV) of the 37 MSPs in our sample as determined using the 3HWC Survey Tool, and as a function of the spindown power, $\dot{E}/d^2$. For the 32 of these pulsars without any significant detection (those with $(\mathrm{TS}{)}^{1/2}$ less than 2), we show the $2\sigma$ upper limit on the flux (blue). For each of the five sources with $(\mathrm{TS}{)}^{1/2}>2$, we plot the $1\sigma$ confidence interval on the gamma-ray flux (red). The black line represents the gamma-ray flux predicted under the assumption that each pulsar generates a TeV halo with the same efficiency as Geminga.

Figure 3

The relationship between $(\mathrm{TS}{)}^{1/2}/F_{\gamma }$ and declination as found using the 3HWC Survey Tool on random sky locations (points). From top-to-bottom, the different colors correspond to the results obtained using the point source template, the 0.5° extended template, the 1° extended template, and the 2° extended template. The dashed lines represent our polynomial fits to these results.

Figure 4

The change to the log-likelihood (relative to no gamma-ray emission, ${\eta }_{\mathrm{MSP}}=0$) as a function of the efficiency for very high-energy gamma-ray production, assuming that all 37 MSPs have the same efficiency. The best-fit value of ${\eta }_{\mathrm{MSP}}=0.74{\eta }_{\text{Geminga}}$ improves the fit by $2\mathrm{\Delta }\ln \mathcal{L}=17.97$, corresponding to a statistical preference of $4.24\sigma$. At the $2\sigma$ level, our fit prefers ${\eta }_{\mathrm{MSP}}=(0.39-1.08)\times {\eta }_{\text{Geminga}}$.

Figure 5

The fraction of 10,000 randomly selected control group realizations that yield evidence for very high-energy gamma-ray emission at a given statistical significance (solid curve). This is compared to the predictions expected from a normal distribution, labeled ${\chi }^2/2$. From among this collection of 10,000 control group samples, only 94 (0.94%) favored a nonzero gamma-ray flux with as much statistical significance ($2\mathrm{\Delta }\ln \mathcal{L}=17.97$) as the collection of 37 MSPs considered in this study.