Updated search for spectral lines from Galactic dark matter interactions with pass 8 data from the Fermi Large Area Telescope

Dark matter in the Milky Way may annihilate directly into $\gamma$ rays, producing a monoenergetic spectral line. Therefore, detecting such a signature would be strong evidence for dark matter annihilation or decay. We search for spectral lines in the Fermi Large Area Telescope observations of the Milky Way halo in the energy range 200 MeV–500 GeV using analysis methods from our most recent line searches. The main improvements relative to previous works are our use of 5.8 years of data reprocessed with the Pass 8 event-level analysis and the additional data resulting from the modified observing strategy designed to increase exposure of the Galactic center region. We search in five sky regions selected to optimize sensitivity to different theoretically motivated dark matter scenarios and find no significant detections. In addition to presenting the results from our search for lines, we also investigate the previously reported tentative detection of a line at 133 GeV using the new Pass 8 data.

Figure 1

P8_CLEAN_V5 energy dispersion at 100 GeV (left panel) and energy resolution as a function of energy (right panel) for the four event types. $E$ is the true energy of the generated $\gamma$ ray and $E^{\prime }$ is the reconstructed energy.

Figure 2

Normalized energy dispersion in the range $5<\log (E)/\mathrm{MeV}<5.2$ from Monte Carlo (MC) with CLEAN event class selection for Pass 8 and Pass 7REP respectively. $E$ is the true energy of the generated $\gamma$ ray and $E^{\prime }$ is the reconstructed energy.

Figure 3

The four DM profiles as a function of the distance from the GC, $r$.

Figure 4

Counts map for the Celestial data set binned in $1°\times 1°$ spatial bins in the R180 ROI, and plotted in Galactic coordinates using the Hammer-Aitoff projection. The energy range is 1–750 GeV. Also shown are the outlines of the other ROIs (R3, R16, R41, and R90) used in this search. The GP region with longitude greater than 6° from the GC and latitude smaller than 5° is removed from all signal ROIs.

Figure 5

Derivation of the global significance for a given local significance in this line search based on 1000 background-only pseudoexperiments of our full fit range. (Left panel) Distribution of the largest $s_{\text{local}}$ value ($s_{\text{max}}$) obtained in each pseudoexperiment. Each pseudoexperiment consisted of 605 fits. (Right panel) Corresponding $s_{\text{local}}$ to $s_{\text{global}}$ conversion.

Figure 6

Fractional signals ($f$) in the GP. Dots show observed $f$ in 31 $10°\times 10°$ boxes along the GP. The solid red line is the average of the statistical uncertainties of the individual boxes. The blue dashed line is the value we chose to characterize $\delta f_{\text{syst}}$ from modeling biases; see text for details.

Figure 7

Distribution of local significance values for all our fits. Best fit to a one-sided Gaussian is shown.

Figure 8

95% C.L. $⟨\sigma v{⟩}_{\gamma \gamma }$ upper limits for each DM profile considered in the corresponding optimized ROI. The upper left panel is for the NFWc ($\gamma =1.3$) DM profile in the R3 ROI. The discontinuity in the expected and observed limit in this ROI around 1 GeV is the result of using only PSF3-type events. See Sec. 3 for more information. The upper right panel is for the Einasto profile in the R16 ROI. The lower left panel is the NFW DM profile in the R41 ROI, and finally the lower right panel is the isothermal DM profile in the R90 ROI. Yellow (green) bands show the 68% (95%) expected containments derived from 1000 no-DM MC simulations (see Sec. 5b). The black dashed lines show the median expected limits from those simulations. Also shown are the limits obtained in our 3.7-year line search [19] and our 5.2-year line search [22] when the assumed DM profiles were the same.

Figure 9

95% C.L. ${\tau }_{\mathrm{DM}}$ lower limits assuming a NFW profile in R180. Yellow (green) bands show the 68% (95%) expected containments derived from 1000 no-DM MC simulations (see Sec. 5b). The black dashed lines show the median expected limits from those simulations. Also shown are the limits obtained in our 3.7-year line search [19] and our 5.2-year line search [22].

Figure 10

Comparison of events selected in R3 from the full data set (5.8 years) passing the Clean event selections simultaneously in both Pass 7REP and Pass 8. The left panel shows the fractional energy difference between Pass 7REP and Pass 8. The black line shows all events that have a reconstructed energy ($E^{\prime }$) above 20 GeV. The dashed red line indicates events which are with Pass 7REP energies in the range 120–150 GeV. The right panel shows the distribution of the distance from the center of the gap of the calorimeter module to the centroid of the cluster. The black curve shows all events above 20 GeV and the red curve has the subset in the Pass 7REP range 120–150 GeV. These events are separated by those near the edge of the tower ($<60\mathrm{mm}$ from the center of the gap), and those away from the edge of the tower ($<60\mathrm{mm}$ from the center of the gap). The blue dashed line marks the edge of the calorimeter module, which is also illustrated at the top of the figure.

Figure 11

Fit at 133 GeV for a $\gamma$ ray in the 3.7-year Pass 8 data set using the 2D energy dispersion model in R3. The solid curve shows the signal and background fitting procedure described in Sec. 5a. The blue dotted line is the signal model that best fits the data. The gray line, which is mostly hidden by the solid curve, is the best-fit background. The bin size is such that the energy resolution is sampled with three bins.

Figure 12

Fit at 133 GeV for a $\gamma$ ray in a 5.8-year Pass 7REP (left panel) and Pass 8 (right panel) data sets using the 2D energy dispersion model in R3. The solid curve shows the signal and background fitting procedure described in Sec. 5a. The blue dotted line is the signal that best fits the data. The gray line, which is mostly hidden by the solid curve, is the best-fit background. The bin size is such that the energy resolution is sampled with three bins.

Figure 13

Fit at 133 GeV for a $\gamma$ ray in a 5.8-year Pass 7REP (left panel) and Pass 8 (right panel) data sets using the 2D energy dispersion model in the Earth’s Limb. The solid curve shows signal and background fitting procedure described in Sec. 5a. The blue dotted line is the signal that best fits the data. The gray line, which is mostly hidden by the solid curve, is the best-fit background. The bin size is such that the energy resolution is sampled with three bins.

Figure 14

(Right) Fractional signals observed in the Galactic plane (GP) scan (gray dots), from the Vela pulsar (blue squares), and from the Earth’s Limb (red triangles). The error bars for the Vela and Limb points are $\delta f_{\text{stat}}=1/\sqrt{b_{\mathrm{eff}}}$.