NuSTAR tests of sterile-neutrino dark matter: New Galactic bulge observations and combined impact

Brandon M. Roach,1, ∗ Kenny C. Y. Ng,2, † Kerstin Perez,1, ‡ John F. Beacom,3, 4, 5, § Shunsaku Horiuchi,6, ¶ Roman Krivonos,7, ∗∗ and Daniel R. Wik8, †† 1Department of Physics, Massachusetts Institute of Technology, Cambridge, MA 02139, USA 2Department of Particle Physics and Astrophysics, Weizmann Institute of Science, Rehovot, Israel 3Center for Cosmology and AstroParticle Physics (CCAPP), Ohio State University, Columbus, OH 43210, USA 4Department of Physics, Ohio State University, Columbus, OH 43210, USA 5Department of Astronomy, Ohio State University, Columbus, OH 43210, USA 6Center for Neutrino Physics, Department of Physics, Virginia Tech, Blacksburg, VA 24061, USA 7Space Research Institute of the Russian Academy of Sciences (IKI), Moscow 117997, Russia 8Department of Physics and Astronomy, University of Utah, Salt Lake City, UT 84112, USA (Dated: Received 4 October 2019; accepted 2 April 2020; published 8 May 2020) We analyze two dedicated NuSTAR observations with exposure ∼190 ks located ∼10◦ from the Galactic plane, one above and the other below, to search for x-ray lines from the radiative decay of sterile-neutrino dark matter. These fields were chosen to minimize astrophysical x-ray backgrounds while remaining near the densest region of the dark matter halo. We find no evidence of anomalous x-ray lines in the energy range 5–20 keV, corresponding to sterile neutrino masses 10–40 keV. Interpreted in the context of sterile neutrinos produced via neutrino mixing, these observations provide the leading constraints in the mass range 10–12 keV, improving upon previous constraints in this range by a factor ∼2. We also compare our results to Monte Carlo simulations, showing that the fluctuations in our derived limit are not dominated by systematic effects. An updated model of the instrumental background, which is currently under develop- ment, will improve NuSTAR’s sensitivity to anomalous x-ray lines, particularly for energies 3–5 keV.

doi: https://doi.org/10.1103/PhysRevD.101.103011

I. INTRODUCTION the detector, allowing for a rejection of photons from known astrophysical sources. Final states with mono- Multiple lines of cosmological evidence indicate that energetic photons are particularly valuable for indirect 80% of the matter density of the Universe, and 25% DM searches, as they result in line-like signals atop a ∼of its energy density, is nonbaryonic and nonluminous,∼ (usually) smooth continuum background. hence its name, dark matter (DM) [1]. At present, the A popular DM candidate with mχ keV is the ster- effects of DM are only measurable via its gravitational ile neutrino, with models such as the∼ νMSM provid- effects on astronomical scales, ranging from the motions ing explanations for the particle nature of DM, neutrino of galaxies and galaxy clusters to the power spectrum masses, and baryogenesis [9–12]. The radiative decay of of the Cosmic Microwave Background [2–7]. The lack sterile neutrinos via χ ν + γ would produce a mono- of a viable Standard Model candidate for particle DM energetic x-ray photon→ and an active neutrino, each with (hereafter symbolized χ) has led to a plethora of theo- E = mχ / 2 [13–20]. retical models, many of which are also motivated by a Sterile neutrinos may be produced in the early Uni- desire to account for other phenomena not explained by verse via mixing with active neutrinos [21], and this pro- the Standard Model (e.g., baryogenesis, neutrino masses, duction may be resonantly enhanced by primordial lep- the hierarchy problem, etc). ton asymmetry [22]. Considerations from big bang nucle- The techniques of indirect detection use astronomi- osynthesis (BBN) [23–25] provide an upper bound on the cal observations to search for the decay and/or anni- cosmological lepton asymmetry per unit entropy density 6 hilation of DM into Standard Model particles such as L6 10 (nν nν¯)/s 2500, which we translate into ≡ − ≤ 2 arXiv:1908.09037v3 [astro-ph.HE] 8 May 2020 electrons/positrons, (anti)protons/nuclei, neutrinos, and the constraint on the active-sterile mixing angle sin 2θ photons [8]. Because photons are not deflected by as- shown in Fig.1 using the sterile-dm code [26]. We trophysical magnetic fields, it is possible to determine note that these BBN limits are particularly sensitive to their arrival direction within the angular resolution of the treatment of neutrino opacities and the plasma equa- tion of state near the QCD phase transition, with dif- ferent calculations finding different results—for example, the limits shown in Refs. [25, 27] for the same value of L6 ∗ [email protected]; orcid.org/0000-0001-8016-2170 † are nearly an order of magnitude less constraining than [email protected]; orcid.org/0000-0001-8016-2170 those from Ref. [26], which the authors of Refs. [16, 28] ‡ [email protected]; orcid.org/0000-0002-6404-4737 § [email protected]; orcid.org/0000-0002-0005-2631 attribute to differences in the treatment of neutrino opac- ¶ [email protected]; orcid.org/0000-0001-6142-6556 ities in the QCD epoch. (An update to the calculation in ∗∗ [email protected]; orcid.org/0000-0003-2737-5673 Ref. [27] is presented in Ref. [41], though the latter does †† [email protected]; orcid.org/0000-0001-9110-2245 2 not present an updated constraint in the mχ sin 2θ − 2

tion mechanism, with all of the sources of uncertainty discussed in the previous paragraph [55–58]. 10 10 NuSTAR Space-based x-ray observatories such as HEAO-1 [59], Chandra [60, 61], XMM-Newton [59, 62–64], Suzaku

Chandra + XMM [38, 65], Fermi-GBM [39], and INTEGRAL [40, 66] have Suzaku provided the most robust constraints on the χ ν + γ 11 → 10 decay rate for mχ 1–100 keV. The observation of an MW Sat. ' Counts unknown x-ray line at E 3.5 keV (“the 3.5-keV line”) in several analyses [33–35] has' led to much interest, as well 2

2 as many follow-up analyses using different instruments 12

n Fermi-GBM i 10 and astrophysical targets [29, 37, 38, 63, 64, 67–83]. s +INTEGRAL Some suggest that the 3.5-keV line may be a signature of sterile-neutrino DM [84] or other DM candidates [85–

13 89]; alternatively, modeling systematics [69, 71] or novel 10 astrophysical processes [90, 91] may play a role. Future BBN Limit (Resonant Production) high-spectral-resolution x-ray instruments may also be able to investigate the DM hypothesis for the origin of the 3.5-keV signal via velocity spectroscopy [92, 93]. 10 14 6 7 8 910 20 30 40 50 Since its launch in 2012, the NuSTAR observatory, due m [ keV ] to its unique large-angle aperture for unfocused x-rays, has provided the leading constraints on sterile-neutrino DM across the mass range 10–50 keV, leveraging observa- FIG. 1. The combined impact on the νMSM parameter space tions of the Bullet Cluster [29], blank-sky fields [31], the of previous NuSTAR searches [29–32] and this work is indi- Galactic center [30], and the M31 galaxy [32]. In each cated by the green region. This work provides the leading of these cases, the NuSTAR observations were originally constraints in the 10–12 keV mass range, as shown in Fig.5. performed to study non-DM phenomena; therefore, DM The tentative E ' 3.5 keV signal [33–35] is indicated by the red point. Constraints from other x-ray instruments [36–40] searches using these data had to contend with large astro- are shown for comparison. Uncertainties associated with MW physical backgrounds and/or reduced effective areas from satellite counts [28] and BBN [26, 27] are discussed in Sec.I. masking bright point sources in the field of view (FOV). Improving upon these constraints, and extending them to the NuSTAR limit of E = 3 keV (e.g., to test the ten- plane.) This lower bound may evolve as calculations are tative 3.5-keV signal), will therefore require observations refined. with lower astrophysical backgrounds, as well as an im- proved model of the low-energy NuSTAR instrumental An additional indirect constraint on sterile-neutrino background. DM arises from comparing the observed number of Milky In this paper, we present new constraints on the decay Way (MW) satellite galaxies to the results of N-body rate of sterile-neutrino DM particles using two NuSTAR cosmological simulations. Compared to cold DM, warm observations, one 10◦ above and the other 10◦ below DM particles are expected to suppress the matter power the Galactic plane,∼ chosen to minimize astrophysical∼ spectrum at small scales, reducing the number of low- x-ray emission while still remaining near the center of mass DM subhaloes orbiting the Galaxy. In Fig.1, the Galactic DM halo. These are the first NuSTAR we adopt the result of Ref. [28] with N = 47, de- subhalo observations dedicated to DM searches. rived from SDSS data. Though a complete review of subhalo constraints on the properties of particle DM is beyond the scope of this paper, we note several impor- In Sec.II, we describe the data reduction and spectral tant points. First, the Milky Way satellite population modeling of the NuSTAR data, consistently incorporat- may not resemble that of a typical galaxy of its size ing the flux from the focused and unfocused FOVs. In and morphology, and surveys of dwarf galaxies targeting Sec.III, we combine the line flux limits from these new observations to constrain the χ ν + γ decay rate for their stellar content must be corrected for completeness → [42]. To address the former issue, surveys such as Satel- sterile neutrinos in the mass range 10–40 keV, obtaining lites Around Galactic Analogues [43] aim to study the the strongest constraints to date in the 10–12 keV mass satellites of Milky Way analogues in the local Universe. range. We conclude in Sec.IV. Recent gravitational lensing surveys have also provided strong constraints on the properties of low-mass (down 8 II. NUSTAR DATA ANALYSIS to . 10 M ) subhaloes at cosmological redshifts unbi- ased by the haloes’ stellar content [44–54]. In all of these cases, constraining mχ using structure observables—both In this section, we outline the aspects of the NuSTAR simulated and observed—also requires a model of the instrument that are relevant to our DM search, and DM power spectrum, which is affected by its produc- describe NuSTAR’s unique wide-angle aperture for un- 3 focused x-rays (Sec.IIA). After describing the recent reduced the effective area, as well as a large contin- NuSTAR off-plane observations (Sec.IIB) and our treat- uum background from the Galactic ridge x-ray emission ment of the NuSTAR instrument response (Sec.IIC), we (GRXE, see Sec.IID) which was the dominant back- conclude with a discussion of the spectral model we use ground component for E . 20 keV. To combat both to analyze the data (Sec.IID). of these issues, we designed two dedicated NuSTAR ob- servations (see TableI), one 10◦ above the Galactic plane (obsID 40410001002), and∼ the other 10◦ below A. The NuSTAR Instrument (40410002002). The high Galactic latitude of∼ these fields was chosen to minimize the GRXE continuum back- The NuSTAR instrument is more fully described in ground while still remaining near the center of the Galac- Refs. [97–99], with the aspects of the instrument relevant tic DM halo, as well as avoiding known bright x-ray for our search technique described in our previous papers sources near the Galactic plane (see Fig.2). [30, 32]. Here, we summarize several key aspects. The NuSTAR observations described above were car- The NuSTAR instrument contains two identical, inde- ried out in August and October 2018, with an ini- pendent, and co-aligned telescopes, each consisting of a tial unfiltered exposure time of 200 ks (summed over grazing-incidence Pt/C-coated x-ray optics module and both obsIDs and FPMs). Data∼ reduction and anal- a Focal Plane Module (FPM). The FPMs (labeled A ysis are performed using the NuSTAR Data Analy- and B) contain an aperture stop, a 100-µm beryllium sis Software pipeline, nustardas v1.5.1. The flags x-ray window with energy-dependent∼ transmission effi- SAAMODE=OPTIMIZED and TENTACLE=YES are used to re- ciency Be(E), and a solid-state CdZnTe detector array move events coincident with NuSTAR passages through with energyE resolution 0.4 keV for x-rays with energies the South Atlantic Anomaly (SAA), and “bad pixels” ∼ E . 20 keV. Within the telescopes, properly-focused in- (defined in the NuSTAR calibration database) are re- coming x-rays reflect twice off the mirror segments, lead- moved. We observe a faint x-ray point source near the ing to their alternative name of 2-bounce (2b) photons. edge of the 2-bounce FOV in obsID 40410001002, whose Both telescopes share essentially-overlapping 130 130 position is consistent with the chromospherically-active FOVs for focused x-rays with energies between 3–79× keV. stellar binary HD 152178 [100, 101]. This system has The lower limit is primarily set by inactive material on also been detected in x-rays by RXTE [102] and Suzaku the surface of the detector and Be(E) (see Secs.IIC [103]. To eliminate systematic uncertainties associated andIID), whereas the upper limitE is set by the Pt K- with modeling this source’s spectrum, we remove from edge of the mirror materials. The maximum x-ray energy our analysis all x-ray events in a circular region of ra- recorded by the detectors is 160 keV. dius 7500 around the nominal position of the source in ∼ Unlike previous focusing x-ray telescopes such as both FPMs, excluding &80% of the source photons [104]. Chandra or XMM-Newton, the 10-m gap between the (The position of the x-ray source 1RXS J165306.1-263434 NuSTAR optics bench and the focal plane is open to also lies within the 2-bounce FOV of this obsID [102]; the sky, allowing stray photons to strike the detector however, it is sufficiently faint that its NuSTAR spec- array without interacting with the mirror elements or trum is consistent with background, so we do not exclude being blocked by the aperture stops. For this reason, it from the analysis. There are no x-ray point sources these unfocused x-rays are called 0-bounce (0b) photons. visible in obsID 40410002002.) Finally, we inspect the Although the 0-bounce effective area A0b is limited by 3–10 keV light-curves of each observation to check for the physical 13 cm2 area of each detector array, the ef- transient fluctuations due to solar activity or unfiltered ∼ fective 0-bounce FOV ∆Ω0b subtended by each array is SAA events, and remove any time intervals with a count 4.5 deg2, nearly two orders of magnitude larger than rate >2.5σ from the quiescent average. After all cuts, the ∼ the 2-bounce FOV ∆Ω2b, and more than counterbalanc- total cleaned exposure time used in this analysis, summed ing the factor of 20 reduction in effective area between over both obsIDs and telescopes, is 190 ks. ∼ ∼ the 2-bounce and 0-bounce apertures. This approach We extract spectra from the full detector planes as ex- provides a large increase in sensitivity to diffuse x-ray tended sources using the nuproducts routine in nus- emission such as that expected from decaying DM in tardas, and bin each spectrum with equal logarith- galactic halos, and thus the 0-bounce technique has been mic separations ∆ log10 E = 0.01 (i.e., 100 bins per the dominant contribution to recent NuSTAR sterile- decade) in the energy ranges 5–20 keV and 95–110 keV. neutrino constraints [30–32]. This provides a statistical uncertainty that is every- where 10% per bin while also being narrower than the 0.4-keV∼ NuSTAR energy resolution across the energy B. NuSTAR Faint-Sky Off-Plane Observations range∼ 5–20 keV. As described in Ref. [32], we exclude the energy range 3–5 keV, as the behavior of the low-energy The previous NuSTAR sterile-neutrino search in the NuSTAR background—particularly the origin of the 3.5- Galactic center region [30] was hampered by the presence and 4.5-keV lines in the default background model—is of bright x-ray point sources in both the 0-bounce and the subject of active investigation. (Additionally, includ- 2-bounce FOVs, whose removal from the data greatly ing the 3–5 keV region can bias the determination of the 4

TABLE I. NuSTAR Galactic Bulge observations used in this analysis, with 0-bounce effective areas after data cleaning. a b c NuSTAR obsID Pointing (J2000) Effective Exposure Detector Area A0b Solid Angle ∆Ω0b RA, Dec (deg) FPMA / B (ks) FPMA / B (cm2) FPMA / B (deg2) 40410001002 253.2508, -26.6472 50.0 / 49.8 11.97 / 11.88 4.36 / 4.62 40410002002 280.3521, -27.6344 44.7 / 44.6 12.71 / 12.60 4.53 / 4.56 a After OPTIMIZED SAA filtering and manual data screening. b After bad pixel removal (both obsIDs) and point-source masking (40410001002 only). c Average solid angle of sky for detecting 0-bounce photons, after correcting for bad pixel removal and vignetting efficiency.

Gal. Longitude 330° 345° 0° 15° 30°

10° b

e d u t i

t 0° a L

. l a G -10°

0.5 0.0 0.5 log10(Flux/mCrab, INTEGRAL 17-60 keV)

FIG. 2. Sky map of the Galactic bulge region. The base color map shows the 17–60 keV flux measured by INTEGRAL [94], with many x-ray point sources clearly visible. The 0-bounce FOVs for the observations analyzed in this paper are indicated by the dashed red (FPMA) and solid blue (FPMB) “Pac-Man”-shaped curves, and avoid known bright x-ray sources. The solid black contours indicate the predicted GRXE flux using the Galactic stellar mass model from Ref. [95] and the GRXE emissivity model from Ref. [96] (see Sec.IID). The contour values are symmetric about b = 0◦, decrease as |b| increases, and are evenly −12.5 −11 −1 −2 −2 spaced in log10(flux) between 10 –10 erg s cm deg , inclusive. internal power-law parameters discussed in Sec.IID; see C. NuSTAR Response Files Ref. [32] for details.) We also exclude the energy range 20–95 keV, as this region is dominated by a forest of in- To describe the effects of the detector effective area strumental lines. DM constraints in this energy range and solid angle for the CXB, GRXE, and DM line com- are therefore weakened and prone to systematic effects, ponents described in Sec.IID, we define custom response as discussed in Refs. [30–32]. Excluding this energy range files that relate the measured event rate d2N/dEdt to also speeds up our analysis, and we verify that it does not the astrophysical flux. For 0-bounce components, the affect our results in the 5–20 keV energy range. Finally, response is Be(E)A0b∆Ω0b, where the grasp A0b∆Ω0b we note that the 20–95 keV energy range has already been E is calculated using the nuskybgd code [98] and Be(E) largely excluded by previous sterile-neutrino searches us- is the Be window transmission efficiency. For 2-bounceE ing data from Fermi-GBM [39], INTEGRAL [40], and components, the response is Be(E)A2b(E)∆Ω2b, where NuSTAR [30–32]. E Be(E) and A2b(E) are calculated by nustardas, ex- E 5 tracting the entire FOV as an extended source using the power-law spectral index and normalization with re- nuproducts. Here, ∆Ω2b is simply the geometric area spect to the internal continuum. As the Earth completely of the 2-bounce FOV, and is 0.046 deg2 for obsID fills the 0-bounce and 2-bounce apertures during occulta- 40410001002 and 0.047 deg2 for obsID∼ 40410002002, the tion mode, we assume that the astrophysical components former being slightly∼ less than the latter due to the ex- contribute negligible flux, and include only the internal clusion of the 7500-radius circle around the point source. detector components when modeling the occulted data. The responses for internal detector components—the in- In particular, the Earth albedo flux is suppressed by at ternal continuum, power-law, and lines—are calculated least one order of magnitude compared to the CXB and by nuproducts, and do not depend on area or solid GRXE over the energy range of our analysis [111–113]. angle. The spectral index and relative normalization of the in- ternal power-law are frozen to their best-fit occultation- mode values during fits to the science data, shown in D. NuSTAR Spectral Modeling Figs.3 and4. This procedure provides a much better 2 fit (χ /d.o.f. . 1.4) to the observed science-mode spec- tra over the energy range of our analysis; however, there Our spectral model contains six components, which are still noticeable deviations, which will be discussed may be broadly classified as having instrumental or as- later in this section, and in Sec.IIIB. trophysical origins (see TableII). The instrumental back- ground consists of a low-energy internal power-law domi- The cosmic x-ray background (CXB) arises from un- nant at energies E . 10 keV, the internal detector contin- resolved extragalactic sources, and constitutes one of the uum, and a series of phenomenologically-motivated lines. dominant irreducible NuSTAR backgrounds in both the The astrophysical components include the cosmic x-ray 0-bounce and 2-bounce FOVs. As specified in the default background (CXB), with an event rate similar to the NuSTAR background model, we parameterize the CXB instrumental components’ over the energy range of this spectrum with a cut-off power-law whose flux, spectral analysis; and the GRXE, whose flux is a factor 10 lower index, and e-folding energy are fixed to the values mea- than the CXB. The treatment of each of these∼ model sured in similar energy ranges by HEAO-1 and INTE- components is described in this section. GRAL [105, 106]; i.e., there are no free parameters in To describe the internal continuum and line back- the CXB model. This choice is supported by a previous grounds, we adopt the default NuSTAR spectral model NuSTAR analysis using the 0-bounce technique, which of Ref. [98]. The internal continuum is parameterized by obtained a CXB flux consistent with our adopted value [109]. We test the effect of allowing the CXB flux to vary a broken power-law with Ebreak = 124 keV, and the line energies and widths are frozen to the values in the default by 10% to account for cross-calibration uncertainty or ± 2 model, with only the line normalizations free to fit. (The the effects of cosmic variance in the 4.5 deg FOV, as ∼ 124-keV break is outside the energy range of our analysis, the number density of CXB sources was previously mea- −2 and thus does not affect the fit; we include it merely for sured by NuSTAR to be &100 deg [114]. We find no continuity with the default NuSTAR model.) The line significant change in the fit quality. Similarly, we ex- normalizations are also allowed to vary between each of amine the effects of allowing the CXB spectral index to the spectra, accounting for differences in the instrumen- be unconstrained. In three of the spectra the best-fit tal background conditions between the FPMs. We retain CXB spectral index is consistent with our adopted value the 95–110 keV data as the event rate in this range is at >90% confidence, whereas in spectrum 40410001002A dominated by the internal continuum, and is necessary the best-fit value is < 1 (inconsistent with previous mea- to constrain the overall continuum normalization. We surements by HEAO-1 and INTEGRAL). The fit quality explore alternative high-energy intervals with endpoints is not significantly improved by allowing the CXB spec- around 95 keV and 120 keV, and find that the fit qual- tral index to vary in any of our spectra, so we fix it to ity is not sensitive to the precise values of the endpoints, the value in TableII. Finally, the highecut term brings a provided the interval is sufficiently wide to constrain the factor exp[(Ecut E)/Efold] for E Ecut and is constant − ≥ −4 internal continuum. for E Ecut, so we choose Ecut = 10 keV to ensure ≤ The default NuSTAR instrumental background model that the exponential folding is applied over the full en- [98] includes a 1-keV collisionally-ionized plasma com- ergy range of our analysis. As shown in Figs.3 and4, the ponent (the apec∼ model in xspec [110]) which is CXB is the dominant astrophysical background in these strongest for energies E < 5 keV and is believed to result off-plane observations. from reflected solar x-rays. Unfortunately, this model The GRXE is believed to result from unresolved point 2 provides a poor fit (χ /d.o.f. & 1.7) to the observed sources in the Galactic ridge [115], and its emissivity spectrum, with the residuals indicating a clear excess in is observed to trace the near-infrared surface brightness the energy range 5–10 keV. As we exclude the E < 5 keV (and hence stellar density) of the Galaxy [96, 108, 116, data, we adopt the procedure described in Refs. [32, 109] 117]. Broadband studies of the GRXE indicate that it and replace the apec model with a power-law. For each is likely a multi-temperature plasma, with kT1 . 1 keV FPM and each obsID, we use the data collected when the and kT2 8 keV [107, 108]. We model the GRXE, telescope aperture is occulted by the Earth to constrain which appears∼ in both the 0-bounce and 2-bounce FOVs, 6

FPMA FPMB 105 2 2

)] χ /60 = 1.30 χ /60 = 1.07 1

− p = 0.06 p = 0.33 104 (cts keV × 103 Line complex region (20–95 keV, excluded) Line complex region (20–95 keV, excluded)

[ keV CXB (total) CXB (total) dE dN Instrumental Instrumental 2 E 10 GRXE (total) GRXE (total)

6 100 2 101 3 101 6 100 1 2 101 3 101 1.2 × × × × 10 × × 1.0 0.8

Data / Model 0.6 10 20 100 10 20 100 E [ keV ]

FIG. 3. Data and model spectra for obsID 40410001002, with FPMA (left) and FPMB (right), including contributions from the CXB, instrumental background, and the GRXE. The error bars correspond to ±1σ statistical uncertainties, and the CXB and GRXE curves incorporate both 0-bounce and 2-bounce emission. We exclude the energy range 20–95 keV as it is dominated by internal detector lines (in previous analyses [30, 32], we have already probed this range well), though we include the energy range 95–110 keV to constrain the internal detector continuum. See Sec.IID for details.

FPMA FPMB 105 2 2

)] χ /60 = 1.01 χ /60 = 1.39 1

− p = 0.45 p = 0.03 104 (cts keV × 103 Line complex region (20–95 keV, excluded) Line complex region (20–95 keV, excluded)

[ keV CXB (total) CXB (total) dE dN Instrumental Instrumental 2 E 10 GRXE (total) GRXE (total)

6 100 2 101 3 101 6 100 1 2 101 3 101 1.2 × × × × 10 × × 1.0 0.8

Data / Model 0.6 10 20 100 10 20 100 E [ keV ]

FIG. 4. Same as Fig.3, but for obsID 40410002002. 7

TABLE II. The NuSTAR spectral model used in this paper. Parameters with numerical values are frozen to those values, and all free parameters are allowed to vary independently between FPMA/B and between the two obsIDs.

Model component xspec modela Parameter Value

CXB powerlaw*highecut 3–20 keV flux 2.6 × 10−11 erg s−1 cm−2 deg−2 [105, 106] Spectral index Γ 1.29 [105, 106] −4 Ecut 10 keV Efold 40 keV [105, 106]

GRXE apec 3–20 keV flux Free Plasma kT 8 keV [107–109] Abundance ratio Free within 0–1.2

Internal continuum bknpower Ebreak 124 keV [98] Γ(E < Ebreak) −0.05 [98] Γ(E > Ebreak) −0.85 [98] Normalization Free

Internal power-law powerlaw Spectral index Γ Frozen for each FPM/obsID (Sec.IID) Relative norm. Frozen for each FPM/obsID (Sec.IID)

Internal lines lorentz Line energies 10.2, 19.7, 104.5 keV [98] Line widths 0.6, 0.2, 0.5 keV [98] Line norms. Free

DM line gaussian Line energy See Sec.IID Line width 0 keV Line flux See Sec.IID a The CXB, GRXE, and DM line models also include absorption from the interstellar medium through the tbabs model with fixed column density NH, as well as absorption from the beryllium x-ray shield. All model components except the internal continuum include the absorption effects of detector surface material. See Sec.IID for details. as a single-temperature collisionally-ionized plasma (the same flux and abundance ratio. apec model described previously) with a fixed tempera- The lower bound on the GRXE abundance ratio arises ture of 8 keV previously measured by NuSTAR; however, from the requirement that elemental abundances be this analysis was not sensitive to the elemental abun- strictly positive, and the upper bound is motivated by dances [109]. (We are unable to leave the GRXE tem- previous measurements of the GRXE [108]. Addition- perature free to fit, as we find that doing so leaves the ally, freezing the abundance ratio to a nonzero value can temperature almost completely unconstrained.) Partic- force the GRXE flux to unreasonable extremes as the ularly strong emission lines between 6–7 keV arise from model attempts to fit the GRXE by way of its emission Kα transitions in neutral and highly-ionized Fe, and it lines, thereby biasing the rest of the 5–20 keV fit. The was these lines which limited the sensitivity of the pre- flux of the GRXE emission lines is directly related to the vious NuSTAR sterile-neutrino search near the Galactic number of atoms in the FOV undergoing electronic de- center (see Ref. [30] and Fig.5 of this paper). excitation, and hence to the elemental abundances of the plasma; as shown by the slight bump in Figs.3 and4, the It is important to note that the “GRXE” component fits to the FPMB spectra of both obsIDs prefer a slightly in our spectral model includes flux from the GRXE, higher GRXE abundance ratio than the FPMA spectra, un-modeled point sources, reflected x-rays from the though this difference is within the uncertainty on the Earth’s atmosphere, and any low-energy instrumental value of the abundance parameter. backgrounds not described by our default spectral model, as the GRXE component includes the only free normal- Finally, the freedom in the GRXE flux acts to account ization parameter in the low-energy part of our spectral for any un-modeled CXB flux, as the two components model. Therefore, we leave both the GRXE elemental have similar continuum shapes in the E < 10 keV range, abundance (as a ratio to solar) and flux as free parame- where their flux is highest. By fixing the CXB and allow- ters, where the flux is unconstrained and the abundance ing the GRXE flux to float, we consistently account for ratio is constrained to the range 0–1.2. The 0-bounce and any variance in the flux of both components, and we find 2-bounce GRXE components are constrained to have the that the best-fit GRXE flux is consistent with Galactic 8 stellar mass and emissivity models [95, 115]. Addition- regions are excesses with respect to the default back- ally, we find that allowing both the CXB and GRXE ground model, the DM line flux limits in the mass ranges fluxes to vary leads to best-fit values which are inconsis- mχ 16 keV and 30–40 keV are correspondingly weak- tent with the previously-described measurements of these ened' (see Sec.IIIB), as we use a conservative line-search components’ flux levels. procedure in which the DM line flux is allowed to fill the We parameterize our DM line signal in xspec with a excess (see Sec.IIIA). In Sec.IIIB, we perform Monte vanishingly-narrow Gaussian—i.e., a δ-function in E— Carlo simulations to verify that our constraint is con- as the intrinsic width of any DM line is expected to be sistent with one limited by statistical variations in our much less than the 0.4 keV detector energy resolution measurement, not systematic variations due to incom- with which it is convolved.∼ Our treatment of the DM line plete modeling. during the line-search procedure is described further in Sec.IIIA. The fluxes of the astrophysical components in our spec- III. NUSTAR DM ANALYSIS tral model—CXB, GRXE, and DM line—are attenuated by absorption and scattering in the interstellar medium In this section, we describe the procedure used to (ISM). This attenuation is parameterized in terms of search for DM line signals and set upper limits on the the equivalent column density of neutral hydrogen, NH, decay rate of DM to final states including a single mono- via the tbabs model in xspec [118]. We adopt fixed energetic photon (Sec.IIIA), and compare to sensitiv- values of 7.0 1020 cm−2 for obsID 40410001002 and ity estimates from simulations (Sec.IIIB). Finally, we 1.1 1021 cm−×2 for obsID 40410002002 [119, 120]. (Both discuss the implications for sterile-neutrino dark matter × FPMs share the same NH value, which is assumed to (Sec.IIIC). be constant across the 0-bounce and 2-bounce FOVs de- spite the somewhat different sky coverage and values of A. DM Line Search ∆Ω0b from A/B.) This corresponds to an optical depth −2 τ . 10 at E = 5 keV, falling steeply with increasing energy. Although the flux attenuation from the ISM is a Equipped with the spectral model described in . 1% effect across the energy range of this analysis, we Sec.IID, we search for DM line signals in the two obser- include it for consistency. vations. Our search procedure follows closely that from Finally, we consider the absorption of x-rays within Refs. [30, 32], and is briefly described here. the NuSTAR instrument itself. Before incoming astro- We divide the 10–40 keV mass band into bins with physical x-rays (from the CXB, GRXE, or DM) strike equal logarithmic separations ∆ log10 mχ = 0.01 (i.e., 100 the detectors, they must pass through a 100-µm beryl- bins per decade in mχ). At each mass bin, we add a DM ∼ lium shield with transmission efficiency Be(E), rising line with photon energy E = mχ/2 to the model. The from 0.67 at E = 3 keV to 0.92 atE E = 5 keV. number of DM photons in the line for each module and ∼ ∼ (The treatment of Be is discussed further in Sec.IIC.) observation is E An additional absorption effect arises in the detectors Γ themselves. The CdZnTe detectors have a 0.11-µm NDM = TA0b ∆Ω0b (1 + f2b), (1) 4πm J Pt contact coating, as well as a 0.27-µm layer∼ of inac- χ ∼ tive CdZnTe (both varying somewhat between individual where Γ is the decay rate, mχ is the DM mass, T is the detector crystals), through which incoming x-rays must observation time, A0b and ∆Ω0b are the 0-bounce effec- pass [99]. At E = 5 keV, these detector components re- tive area and effective FOV defined in Sec.IIC, is sult in a flux attenuation of 25%, though this decreases the FOV-averaged line-of-sight integral of the DMJ den- ∼ quickly with increasing energy [32]. These detector ab- sity (J-factor), and f2b is the energy-dependent contribu- sorption effects (often called nuabs or detabs) are in- tion from the 2-bounce component (see Fig. 3 in Ref. [32] cluded in every spectral component except the internal for the energy dependence of the 2-bounce contribution; continuum. in this work, we find a modest 20% enhancement at As shown in Figs.3 and4, the model described in E = 10 keV). ∼ Sec.IID provides an acceptable fit to the NuSTAR spec- To obtain the J-factors, we consider several tra across most of the 5–20 keV energy range (see Figs.3 DM density profiles. One popular choice is the and4 for the reduced- χ2 and corresponding p-values for generalized Navarro-Frenk-White (NFW) profile, −γ γ−3 each spectrum), but there are several deviations from ρ (r/rs) (1 + r/rs) . For the standard (DM- the model that may affect our derived line flux lim- only)∝ NFW profile, we adopt an inner slope γ = 1 its, and thus require further consideration. The higher and scale radius rs = 20 kpc [121, 122]. We fix the χ2 in FPMA of obsID 40410001002A is due to the en- galactocentric solar radius and the local DM density ergy range 15–20 keV (excluding this energy range yields to be 8 kpc and 0.4 GeV cm−3, respectively [123– χ2/47 = 0.94 with p = 0.59), and similarly for FPMB of 125]. The standard NFW profile was found to be obsID 40410002002B in the energy range 8–9 keV (yield- a good fit to the Milky Way kinematic data [126], ing χ2/54 = 1.15 with p = 0.21). As both of these but it has been suggested that the density profiles 9

Neronov+ (2016) Observed limit Perez+ (2017) 68% expected Ng+ (2019) 27 27 95% expected 10 This work 10 ]

1 s

[

10 28 10 28

10 29 10 29 6 7 8 910 20 30 40 50 10 15 20 25 30 40 m [ keV ] m [ keV ]

FIG. 5. Left: Comparison of the limit obtained in this paper to that from several surveys using the 0-bounce technique, including blank sky (green, Ref. [31]), Galactic center (red, Ref. [30]), and M31 fields (blue, Ref. [32]), as well as the tentative signal at E ' 3.5 keV (red point, Refs. [33–35]). With only ∼190 ks, we have achieved comparable constraints to analyses with much deeper exposures [30–32]. We have achieved the best constraint in 10–12 keV mass range, essential for investigating the remaining νMSM parameter space shown in Fig.1. Right: The observed 95% upper limit on the DM decay rate Γ obtained in this paper, compared to the expected 68% (green) and 95% (yellow) sensitivity bands from simulations (see Sec.IIIB). could be flattened below 1.5 kpc [127, 128]. There- adding the respective χ2 distributions: fore, we also consider the more conservative choice, coreNFW, where we set a density core below 1.5 kpc— 2 X 2 i.e., ρ(r < 1.5 kpc) = ρ(1.5 kpc). Another conservative X (Γ) = χ (Γ) . (2) NFW variant we consider is the sNFW, where we use obs a shallower index γ = 0.7 [123]. We use sNFW as our default result, obtaining 20 GeV cm−3 kpc sr−1 for We note that for each module, the background param- the observation regions inJ' this analysis; for NFW and eters are allowed to be independent (see Sec.IID for coreNFW, the J-factors are larger by 20%. exceptions). Compared with simply stacking the spec- ∼ tra, this combining procedure is used to avoid potential Another shallow density profile often con- systematic errors due to combining observations with dif- sidered in the literature is the Burkert profile −1 2 2 −1 ferent instrumental and/or astrophysical backgrounds. ρ (1 + r/rs) (1 + r /rs ) [129], with best-fit local∝ DM density 0.5 GeV cm−3 and scale radius The minimum in X2(Γ) for each mass bin corresponds 8 kpc [126, 130]. This∼ profile effectively has a density to the best-fit decay rate Γmin, with a 5σ line detection 2 2 core∼ within r , which we note is much larger than what requiring X (Γmin) X (Γ = 0) < 25. We find no s − − was found in Refs. [127, 128]. Even in this case, the signals consistent with decaying DM in the mass range J-factor is only 10% smaller than our default sNFW 10–40 keV, and instead set upper limits on the DM de- choice. This small∼ deviation shows the robustness of our cay rate. The 95% one-sided upper limit, Γ95, occurs 2 2 results, and reflects an additional advantage of using at X (Γ95) = X (Γmin) + 2.71, and is shown in both observations slightly offset from the Galactic center. frames of Fig.5. In the 10–40 keV mass range, our re- sults are comparable to previous NuSTAR limits from At each DM mass, the only free parameter for the DM 2 blank-sky [31], Galactic center [30], and M31 observa- line is the decay rate. We find the best-fit χ (Γ) dis- tions [32]. In particular, we are able to improve upon tribution for each module and observation by scanning previous constraints in the 10–12 keV mass range by a through a range of Γ, refitting the entire spectral model 2 factor of 2. Finally, we note that with only 190 ks to find the minimum χ value for each Γ. This line-search exposure,∼ our dedicated Galactic bulge observations∼ are procedure is conservative, as it allows the DM line to at- able to achieve sensitivity comparable with searches us- tain the full strength of any background lines. ing several Ms combined exposure. This is due to the low The sensitivity of the two observations (four separate astrophysical background, as well as the large J-factors fits including both modules) at each mχ are combined by in the chosen FOVs. 10

B. Sensitivity Estimation with Simulations excluding these energy ranges from our analysis a poste- riori. To validate our results, we perform line searches in We also test the effect of incorporating a flat 7.5% sys- mock spectra to find the expected upper limits when the tematic across the entire energy range of all four spec- 2 spectra are purely statistically limited. This exercise also tra, sufficient to give χ /60 . 1 for each. We run these allows us to further study the deviations discussed in spectra through the same line-search procedure as our Sec.IID. default analysis, and find a combined DM limit that is Instead of fully mimicking the actual analysis, where a factor 1.5 weaker than our default result, but still a ∼ we analyze each module separately and then combine the factor 1.5 stronger than the previously-leading Ref. [31] ∼ constraints, we simplify the procedure by considering a in the mass range 10–12 keV. We conclude that any sys- single spectrum (rather than all four) per mock analysis tematic effects on our final DM limit are subdominant to speed up the computation. We generate 100 Monte to the range expected from statistical fluctuations al- Carlo (MC) spectra with no DM line, using the fakeit ready shown in the right panel of Fig.5. Adding such tool in xspec. Each spectrum has 200 ks exposure, and a flat systematic to all four spectra is also an extreme is generated using the best-fit spectral model of FPMA, procedure considering the two spectra lacking these ex- obsID 40410001002. This simplification is motivated by cesses (40410001002B and 40410002002A), as well as the the fact that the spectrum for each module has similar isolated energy ranges in which these excesses appear; best-fit model parameters, and hence statistics. We also therefore, we do not apply such a flat systematic when test the results obtained with 10 of these simplified sim- calculating our default DM limit. ulations against 10 full realizations (i.e., including both It is plausible that the excesses described previously obsIDs and both FPMs) and find good agreement. We result from some un-modeled, transient background com- then pass these mock spectra through the same fitting ponent. Such a component was not evident during our and line-search procedure as the data. At each mass bin, initial data screening (see Sec.IIB), but there known we thus have 100 simulated upper limits. We interpo- issues with the default NuSTAR background model in late the cumulative distribution of these upper limits and these regions. (As noted previously, the excesses in the 8– find the corresponding 68% and 95% intervals. The up- 9 keV and 15–20 keV energy ranges are inconsistent with per limits can then obtained directly from the line-search DM.) If we were to add additional background compo- procedure (see Sec.IIIA) without needing to combine nents in these regions, our DM limit in those regions different FPMs. would become stronger, as some of the flux assigned to The right panel of Fig.5 shows the expected upper the DM line would instead be incorporated into the new limit bands obtained with the mock spectra. Our upper background components. Elucidating the form of these limits obtained from real data are consistent with the additional background components—if they exist—is be- MC expectation across most of the 10–40 keV mass range yond the scope of this work, and will require analysis at the 2σ level; however, there are several features that (ongoing) of NuSTAR datasets with significantly longer warrant closer attention. exposure time. As described in Sec.IID, the high χ2 values for spec- We conclude by considering the ranges where the DM tra 40410001002A and 40410002002B are caused primar- limit in Fig.5 most departs from the MC expectation, ily by isolated excesses in the energy ranges 15–20 keV though in all cases the observed limit remains consistent and 8–9 keV, respectively. We first consider∼ the possi- with the 95% MC band. (The upward fluctuations in the bility∼ of these excesses being purely statistical. Though observed limit near masses 16–18 keV and 30–40 keV the corresponding p-values are small—0.06 and 0.03, have already been discussed.)∼ First, the upward∼ fluctu- respectively—this possibility is supported by these ex- ations near the edges of the region of interest (masses cesses appearing in only two of the spectra, and in two 10 keV and 40 keV) likely arise from parts of the DM different energy ranges. Additionally, the upward fluc- line leaving the energy range 5–20 keV. Second, the up- tuations in the observed limit lie within the 95% band ward fluctuation in the MC band near mχ 20 keV expected from MC simulations incorporating only statis- is attributed to a weak line near E 10 keV' in the tical fluctuations (see the right panel of Fig.5). If we background model, whereas the observed' limit exhibits a consider the extreme procedure of excluding the energy downward fluctuation due to negative residuals in spec- ranges 15–20 keV and 8–9 keV in spectra 40410001002A trum 40410001002B at E 10 keV. Finally, we turn and 40410002002B, respectively, the DM limits in the to the mass range 10–12' keV, where our results im- mass ranges 16–18 keV and 30–40 keV are strength- prove the most compared∼ to previous analyses and the ened by a factor∼ 1.3, as we are∼ no longer including data observed limit also touches the lower end of the MC band. which favor nonzero∼ DM flux; excluding these excesses A closer inspection shows that this is driven by several also reduces the best-fit continuum level over the rest downward-fluctuating data points from 40410001002A of the energy range, slightly weakening the overall limit and 40410002002A/B. These negative residuals appear elsewhere by at most a factor 1.5. In both cases, the at different energies in three different modules, and the changes in the DM limit are well∼ within the MC band bin widths are a factor 4 narrower than the detector of Fig.5, so we do not pursue the extreme procedure of energy resolution. This lends∼ support to the strong limit 11 being caused by statistical downward fluctuations. from the radiative decay of sterile-neutrino DM in the Galactic halo. The observation regions were optimized to reduce astrophysical x-ray backgrounds from Galac- C. Sterile-Neutrino DM Constraints tic x-ray sources and from the Galactic ridge x-ray emis- sion while remaining near the center of the Galactic halo, For sterile-neutrino DM, we convert the decay rate con- where the DM decay signal is expected to be strongest. straints to mixing angle constraints using [17, 18] We consistently model the flux from both the focused (2- bounce) and unfocused (0-bounce) NuSTAR apertures,  2  5 though our sensitivity to decaying DM is dominated by −32 −1 sin 2θ  mχ  Γ = 1.38 10 s . (3) the large unfocused FOV. To avoid the systematic effects × 10−10 keV of stacking spectra with different instrumental and astro- The aggregate constraints in the mass-mixing-angle plane physical backgrounds, we model the spectra individually from x-ray searches (including NuSTAR) are shown in and combine the sensitivity of each. Fig.1. As described previously, our high-latitude Galac- Finding no evidence of sterile-neutrino DM decays, we tic bulge constraints are a factor 2 stronger than the instead set upper limits on the sterile neutrino decay previous leading limits [31] in the mass∼ range 10–12 keV rate in the mass range 10–40 keV. In the mass range while requiring a factor 50 less exposure time, and are 10–12 keV, our limits are a factor 2 stronger than ∼ ∼ comparable with previous∼ NuSTAR constraints over the the previous leading limits while requiring a factor 50 ∼ rest of the 10–40 keV mass range. This supports the use less exposure time. This is due in part to the low as- of observation regions with low astrophysical background trophysical background and large J-factor in these opti- and large J-factors. mized observation regions, as well as downward statistical In the context of the νMSM, the parameter space is fluctuations. We also perform Monte Carlo simulations also bounded by production and structure formation con- to determine our expected DM sensitivity, and find that straits [26, 28] (see also Ref. [32] for discussion). As dis- our derived limits are consistent with expectations across cussed in Sec.IIIB, the DM line analysis in this paper most of the 10–40 keV mass range. is limited mostly by statistics, except for the known fea- As the astrophysical background (now dominated by ture near E 15 keV. To cover the νMSM window for the irreducible CXB flux) in these observations is com- ' mχ > 10 keV, a factor 4 improvement in sensitivity is parable to the instrumental background, we observe needed, corresponding to∼ 4 Ms exposure of regions with deviations of the spectra from the default NuSTAR large J-factors and minimal∼ astrophysical backgrounds background model, particularly in the energy ranges (similar to the present paper). Though a survey of this 8–9 keV and 15–20 keV. Though similar effects are ∼ ∼ depth is feasible, we caution that systematic deviations visible in other NuSTAR analyses (see the left panel of from the default NuSTAR background model will likely Fig.5), the excesses in our spectra are consistent with prevent long exposures from reaching their design sensi- statistical fluctuations (see Sec.IIIB). Detailed charac- tivity until an improved model of the NuSTAR instru- terization of the instrumental background is ongoing, and mental background can be developed. Ongoing work for additional NuSTAR searches, particularly with an im- improving the NuSTAR instrumental background model, proved model of the instrumental background, will be especially in the 3–5 keV energy range, will be essential uniquely suited to probing the remaining νMSM param- for further testing of the νMSM down to mχ = 6 keV, eter space, as well as investigating the nature of the including the tentative signal at E 3.5 keV. 3.5-keV line. '

IV. CONCLUSIONS AND OUTLOOK ACKNOWLEDGMENTS

The NuSTAR observatory’s large FOV for unfocused We thank Alexey Boyarsky, Steve Rossland, Oleg x-rays has been pivotal in constraining the proper- Ruchayskiy, and Shuo Zhang for helpful comments and ties of sterile-neutrino DM with mχ keV, such as discussions. We also thank the anonymous referees for that predicted by the νMSM. NuSTAR∼ observations their constructive comments. of the Galactic center, blank-sky extragalactic fields, The NuSTAR observations described in this work were and M31 have provided world-leading constraints on awarded under NASA Grant No. 80NSSC18K1615. We the χ ν + γ decay rate in the mass range 10–50 keV, thank the NuSTAR team at NASA, JPL, and CalTech practically→ closing the “window” in the νMSM parame- for the excellent performance of the instrument and their ter space for masses 20–50 keV. Closing the window for assistance with initial data processing. masses 6–20 keV, however, has proved difficult, due to The computational aspects of this work made exten- large astrophysical x-ray backgrounds in the observation sive use of the following packages: saoimage ds9 dis- regions. tributed by the Smithsonian Astrophysical Observatory; In this paper, we analyze a combined 190 ks of the scipy ecosystem [131], particularly matplotlib and NuSTAR observations to search for x-rays∼ originating numpy; and astropy, a community-developed core 12 python package for Astronomy [132, 133]. This re- fellowships. K.C.Y.N. is supported by a Croucher Fel- search has made use of data and software provided by the lowship and a Benoziyo Fellowship. K.P. receives addi- High Energy Astrophysics Science Archive Research Cen- tional support from the Alfred P. Sloan Foundation and ter (HEASARC), which is a service of the Astrophysics RCSA Cottrell Scholar Award No. 25928. J.F.B. is sup- Science Division at NASA/GSFC and the High Energy ported by NSF Grant No. PHY-1714479. S.H. is sup- Astrophysics Division of the Smithsonian Astrophysical ported by the U.S. Department of Energy under Award Observatory. No. DE-SC0018327, as well as NSF Grants No. AST- 1908960 and PHY-1914409. R.K. receives support from B.M.R. and K.P. receive support from NASA Grant the Russian Science Foundation under Grant No. 19-12- No. 80NSSC18K1615. B.M.R. is also partially supported 00396. D.R.W. is supported by NASA ADAP Grant No. by MIT Department of Physics and School of Science 80NSSC18K0686.

[1] M. Tanabashi et al. (Particle Data Group), “Review of [16] A. Boyarsky, M. Drewes, T. Lasserre, S. Mertens, and Particle Physics,” Phys. Rev. D 98, 030001 (2018). O. Ruchayskiy, “Sterile Neutrino Dark Matter,” Prog. [2] G. Bertone, D. Hooper, and J. Silk, “Particle Dark Part. Nucl. Phys. 104, 1 (2019). 0 Matter: Evidence, Candidates and Constraints,” Phys. [17] R. Shrock, “Decay L → νlγ in Gauge Theories of Weak Rept. 405, 279 (2005), arXiv:hep-ph/0404175 [hep-ph]. and Electromagnetic Interactions,” Phys. Rev. D9, 743 [3] L. E. Strigari, “Galactic Searches for Dark Matter,” (1974). Phys. Rept. 531, 1 (2013), arXiv:1211.7090 [astro- [18] P. B. Pal and L. Wolfenstein, “Radiative Decays of Mas- ph.CO]. sive Neutrinos,” Phys. Rev. D25, 766 (1982). [4] M. S. Seigar, Dark Matter in the Universe (Morgan & [19] A. D. Dolgov and S. H. Hansen, “Massive Sterile Neu- Claypool Publishers, 2015). trinos as Warm Dark Matter,” Astroparticle Phys. 16, [5] M. R. Buckley and A. H. G. Peter, “Gravitational 339 (2002), arXiv:hep-ph/0009083 [hep-ph]. Probes of Dark Matter Physics,” Phys. Rept. 761, 1 [20] K. Abazajian, G. M. Fuller, and W. H. Tucker, “Di- (2018), arXiv:1712.06615 [astro-ph.CO]. rect Detection of Warm Dark Matter in the X-Ray,” [6] N. Aghanim et al. (Planck Collaboration), “Planck Astrophys. J. 562, 593 (2001), arXiv:astro-ph/0106002 2018 Results. VI. Cosmological Parameters,” (2018), [astro-ph]. arXiv:1807.06209 [astro-ph.CO]. [21] S. Dodelson and L. M. Widrow, “Sterile Neutrinos [7] R. H. Wechsler and J. L. Tinker, “The Connec- as Dark Matter,” Phys. Rev. Lett. 72, 17 (1994), tion Between Galaxies and Their Dark Matter Ha- arXiv:hep-ph/9303287 [hep-ph]. los,” Ann. Rev. Astron. Astrophys. 56, 435 (2018), [22] X.-D. Shi and G. M. Fuller, “A New Dark Matter Can- arXiv:1804.03097 [astro-ph.GA]. didate: Nonthermal Sterile Neutrinos,” Phys. Rev. Lett. [8] J. M. Gaskins, “A Review of Indirect Searches for Par- 82, 2832 (1999), arXiv:astro-ph/9810076 [astro-ph]. ticle Dark Matter,” Contemp. Phys. 57, 496 (2016), [23] A. D. Dolgov, S. H. Hansen, S. Pastor, S. T. Pet- arXiv:1604.00014 [astro-ph.HE]. cov, G. G. Raffelt, and D. V. Semikoz, “Cosmological [9] T. Asaka, S. Blanchet, and M. Shaposhnikov, “The Bounds on Neutrino Degeneracy Improved by Flavor νMSM, Dark Matter and Neutrino Masses,” Phys. Lett. Oscillations,” Nucl. Phys. B632, 363 (2002), arXiv:hep- B631, 151 (2005), arXiv:hep-ph/0503065 [hep-ph]. ph/0201287 [hep-ph]. [10] T. Asaka, M. Laine, and M. Shaposhnikov, “Light- [24] P. D. Serpico and G. G. Raffelt, “Lepton Asymme- est Sterile Neutrino Abundance within the νMSM,” try and Primordial Nucleosynthesis in the Era of Pre- J. High Energy Phys. 01, 091 (2007), [Erratum: cision Cosmology,” Phys. Rev. D71, 127301 (2005), JHEP02,028(2015)], arXiv:hep-ph/0612182 [hep-ph]. arXiv:astro-ph/0506162 [astro-ph]. [11] L. Canetti, M. Drewes, T. Frossard, and M. Shaposh- [25] A. Boyarsky, O. Ruchayskiy, and M. Shaposhnikov, nikov, “Dark Matter, Baryogenesis and Neutrino Oscil- “The Role of Sterile Neutrinos in Cosmology and As- lations from Right Handed Neutrinos,” Phys. Rev. D87, trophysics,” Ann. Rev. Nucl. Part. Sci. 59, 191 (2009), 093006 (2013), arXiv:1208.4607 [hep-ph]. arXiv:0901.0011 [hep-ph]. [12] L. Canetti, M. Drewes, and M. Shaposhnikov, “Sterile [26] T. Venumadhav, F.-Y. Cyr-Racine, K. N. Abazajian, Neutrinos as the Origin of Dark and Baryonic Matter,” and C. M. Hirata, “Sterile Neutrino Dark Matter: Weak Phys. Rev. Lett. 110, 061801 (2013), arXiv:1204.3902 Interactions in the Strong Coupling Epoch,” Phys. Rev. [hep-ph]. D94, 043515 (2016), arXiv:1507.06655 [astro-ph.CO]. [13] A. Kusenko, “Sterile Neutrinos: The Dark Side of [27] M. Laine and M. Shaposhnikov, “Sterile Neutrino Dark the Light Fermions,” Phys. Rept. 481, 1 (2009), Matter as a Consequence of νMSM-Induced Lepton arXiv:0906.2968 [hep-ph]. Asymmetry,” J. Cosmology Astroparticle Phys. 806, [14] M. Drewes et al., “A White Paper on keV Sterile Neu- 031 (2008), arXiv:0804.4543 [hep-ph]. trino Dark Matter,” J. Cosmology Astroparticle Phys. [28] J. F. Cherry and S. Horiuchi, “Closing in on Resonantly 1701, 025 (2017), arXiv:1602.04816 [hep-ph]. Produced Sterile Neutrino Dark Matter,” Phys. Rev. [15] K. N. Abazajian, “Sterile Neutrinos in Cosmology,” D95, 083015 (2017), arXiv:1701.07874 [hep-ph]. Phys. Rept. 711-712, 1 (2017), arXiv:1705.01837 [hep- [29] S. Riemer-Sørensen et al., “Dark Matter Line Emis- ph]. sion Constraints from NuSTAR Observations of 13

the Bullet Cluster,” Astrophys. J. 810, 48 (2015), Rev. Lett. 121, 211302 (2018), arXiv:1711.06267 [astro- arXiv:1507.01378 [astro-ph.CO]. ph.CO]. [30] K. Perez, K. C. Y. Ng, J. F. Beacom, C. Hersh, S. Ho- [43] M. Geha et al., “The SAGA Survey. I. Satellite Galaxy riuchi, and R. Krivonos, “Almost Closing the νMSM Populations around Eight Milky Way Analogs,” Astro- Sterile Neutrino Dark Matter Window with NuSTAR,” phys. J. 847, 4 (2017), arXiv:1705.06743 [astro-ph.GA]. Phys. Rev. D95, 123002 (2017), arXiv:1609.00667 [44] L. V. E. Koopmans, “Gravitational Imaging of Cold [astro-ph.HE]. Dark Matter Substructures,” Mon. Not. Roy. Astron. [31] A. Neronov, D. Malyshev, and D. Eckert, “De- Soc. 363, 1136 (2005), arXiv:astro-ph/0501324. caying Dark Matter Search with NuSTAR Deep [45] S. Vegetti and L. V. E. Koopmans, “Bayesian Strong Sky Observations,” Phys. Rev. D94, 123504 (2016), Gravitational-Lens Modelling on Adaptive Grids: Ob- arXiv:1607.07328 [astro-ph.HE]. jective Detection of Mass Substructure in Galax- [32] K. C. Y. Ng, B. M. Roach, K. Perez, J. F. Beacom, ies,” Mon. Not. Roy. Astron. Soc. 392, 945 (2009), S. Horiuchi, R. Krivonos, and D. R. Wik, “New Con- arXiv:0805.0201 [astro-ph]. straints on Sterile Neutrino Dark Matter from NuSTAR [46] S. Vegetti, O. Czoske, and L. V. E. Koopmans, “Quan- M31 Observations,” Phys. Rev. D 99, 083005 (2019), tifying Dwarf Satellites Through Gravitational Imag- arXiv:1901.01262 [astro-ph.HE]. ing: The Case of SDSS J120602.09+514229.5,” Mon. [33] A. Boyarsky, O. Ruchayskiy, D. Iakubovskyi, and Not. Roy. Astron. Soc. 407, 225 (2010), arXiv:1002.4708 J. Franse, “Unidentified Line in X-Ray Spectra of the [astro-ph.CO]. Andromeda Galaxy and Perseus Galaxy Cluster,” Phys. [47] S. Vegetti, L. V. E. Koopmans, A. Bolton, T. Treu, and Rev. Lett. 113, 251301 (2014), arXiv:1402.4119 [astro- R. Gavazzi, “Detection of a Dark Substructure through ph.CO]. Gravitational Imaging,” Mon. Not. Roy. Astron. Soc. [34] E. Bulbul, M. Markevitch, A. Foster, R. K. Smith, 408, 1969 (2010), arXiv:0910.0760 [astro-ph.CO]. M. Loewenstein, and S. W. Randall, “Detection of an [48] S. Vegetti, D. Lagattuta, J. McKean, M. W. Auger, Unidentified Emission Line in the Stacked X-Ray Spec- C. D. Fassnacht, and L. V. E. Koopmans, “Gravita- trum of Galaxy Clusters,” Astrophys. J. 789, 13 (2014), tional Detection of a Low-Mass Dark Satellite Galaxy arXiv:1402.2301 [astro-ph.CO]. at Cosmological Distance,” Nature 481, 341 (2012), [35] A. Boyarsky, J. Franse, D. Iakubovskyi, and arXiv:1201.3643 [astro-ph.CO]. O. Ruchayskiy, “Checking the Dark Matter Origin of [49] S. Vegetti, L. V. E. Koopmans, M. W. Auger, T. Treu, a 3.53 keV Line with the Milky Way Center,” Phys. and A. S. Bolton, “Inference of the Cold Dark Matter Rev. Lett. 115, 161301 (2015), arXiv:1408.2503 [astro- Substructure Mass Function at z = 0.2 Using Strong ph.CO]. Gravitational Lenses,” Mon. Not. Roy. Astron. Soc. [36] S. Horiuchi, B. Bozek, K. N. Abazajian, M. Boylan- 442, 2017 (2014), arXiv:1405.3666 [astro-ph.GA]. Kolchin, J. S. Bullock, S. Garrison-Kimmel, and [50] S. Birrer, A. Amara, and A. Refregier, “Lensing Sub- J. Onorbe, “Properties of Resonantly Produced Ster- structure Quantification in RXJ1131-1231: A 2 keV ile Neutrino Dark Matter Subhaloes,” Mon. Not. Roy. Lower Bound on Dark Matter Thermal Relic Mass,” Astron. Soc. 456, 4346 (2016), arXiv:1512.04548 [astro- J. Cosmology Astroparticle Phys. 2017, 37 (2017), ph.CO]. arXiv:1702.00009 [astro-ph.CO]. [37] O. Ruchayskiy, A. Boyarsky, D. Iakubovskyi, E. Bul- [51] S. Vegetti, G. Despali, M. R. Lovell, and W. Enzi, bul, D. Eckert, J. Franse, D. Malyshev, M. Markevitch, “Constraining Sterile Neutrino Cosmologies with and A. Neronov, “Searching for Decaying Dark Mat- Strong Gravitational Lensing Observations at Redshift ter in Deep XMM–Newton Observation of the Draco z ∼ 0.2,” Mon. Not. Roy. Astron. Soc. 481, 3661 (2018), Dwarf Spheroidal,” Mon. Not. Roy. Astron. Soc. 460, arXiv:1801.01505 [astro-ph.CO]. 1390 (2016), arXiv:1512.07217 [astro-ph.HE]. [52] D. Gilman, S. Birrer, T. Treu, A. Nierenberg, and [38] T. Tamura, R. Iizuka, Y. Maeda, K. Mitsuda, and N. Y. A. Benson, “Probing Dark Matter Structure down to Yamasaki, “An X-Ray Spectroscopic Search for Dark 107 Solar Masses: Flux Ratio Statistics in Gravitational Matter in the Perseus Cluster with Suzaku,” Publ. As- Lenses with Line-of-Sight Haloes,” Mon. Not. Roy. As- tron. Soc. Japan 67, 23 (2015), arXiv:1412.1869 [astro- tron. Soc. 487, 5721 (2019), arXiv:1901.11031 [astro- ph.HE]. ph.CO]. [39] K. C. Y. Ng, S. Horiuchi, J. M. Gaskins, M. Smith, [53] J.-W. Hsueh, W. Enzi, S. Vegetti, M. W. Auger, and R. Preece, “Improved Limits on Sterile Neu- C. D. Fassnacht, G. Despali, L. V. E. Koopmans, and trino Dark Matter using Full-Sky Fermi Gamma-Ray J. P. McKean, “SHARP—VII. New Constraints on the Burst Monitor Data,” Phys. Rev. D92, 043503 (2015), Dark Matter Free-Streaming Properties and Substruc- arXiv:1504.04027 [astro-ph.CO]. ture Abundance from Gravitationally Lensed Quasars,” [40] A. Boyarsky, D. Malyshev, A. Neronov, and Mon. Not. Roy. Astron. Soc. (2019), 10.1093/mn- O. Ruchayskiy, “Constraining Dark Matter Properties ras/stz3177, arXiv:1905.04182 [astro-ph.CO]. with SPI,” Mon. Not. Roy. Astron. Soc. 387, 1345 [54] D. Gilman, S. Birrer, A. Nierenberg, T. Treu, (2008), arXiv:0710.4922 [astro-ph]. X. Du, and A. Benson, “Warm Dark Matter [41] J. Ghiglieri and M. Laine, “Improved Determination of Chills Out: Constraints on the Halo Mass Func- Sterile Neutrino Dark Matter Spectrum,” J. High En- tion and the Free-Streaming Length of Dark Mat- ergy Phys. 11, 171 (2015), arXiv:1506.06752 [hep-ph]. ter with Eight Quadruple-Image Strong Gravitational [42] S. Y. Kim, A. H. G. Peter, and J. R. Hargis, “Miss- Lenses,” Mon. Not. Roy. Astron. Soc. 491, 6077 (2020), ing Satellites Problem: Completeness Corrections to arXiv:1908.06983 [astro-ph.CO]. the Number of Satellite Galaxies in the Milky Way are [55] D. Boyanovsky, “Clustering Properties of a Sterile Neu- Consistent with Cold Dark Matter Predictions,” Phys. trino Dark Matter Candidate,” Phys. Rev. D78, 103505 14

(2008), arXiv:0807.0646 [astro-ph]. Soc. 452, 3905 (2015), arXiv:1408.4115 [astro-ph.HE]. [56] D. Boyanovsky and J. Wu, “Small Scale Aspects [71] O. Urban, N. Werner, S. W. Allen, A. Simionescu, J. S. of Warm Dark Matter: Power Spectra and Acous- Kaastra, and L. E. Strigari, “A Suzaku Search for tic Oscillations,” Phys. Rev. D83, 043524 (2011), Dark Matter Emission Lines in the X-Ray Brightest arXiv:1008.0992 [astro-ph.CO]. Galaxy Clusters,” Mon. Not. Roy. Astron. Soc. 451, [57] J.-L. Kuo, M. Lattanzi, K. Cheung, and J. W. F. 2447 (2015), arXiv:1411.0050 [astro-ph.CO]. Valle, “Decaying Warm Dark Matter and Structure For- [72] N. Sekiya, N. Y. Yamasaki, and K. Mitsuda, “A Search mation,” J. Cosmology Astroparticle Phys. 1812, 026 for a keV Signature of Radiatively Decaying Dark Mat- (2018), arXiv:1803.05650 [astro-ph.CO]. ter with Suzaku XIS Observations of the X-Ray Diffuse [58] N. Menci, A. Grazian, A. Lamastra, F. Calura, Background,” Publ. Astron. Soc. Jap. 68, S31 (2016), M. Castellano, and P. Santini, “Galaxy Formation in arXiv:1504.02826 [astro-ph.HE]. Sterile Neutrino Dark Matter Models,” Astrophys. J. [73] E. Figueroa-Feliciano et al. (XQC Collaboration), 854, 1 (2018), arXiv:1801.03697 [astro-ph.CO]. “Searching for keV Sterile Neutrino Dark Matter [59] A. Boyarsky, A. Neronov, O. Ruchayskiy, and M. Sha- with X-Ray Microcalorimeter Sounding Rockets,” As- poshnikov, “Constraints on Sterile Neutrinos as a trophys. J. 814, 82 (2015), arXiv:1506.05519 [astro- Dark Matter Candidate from the Diffuse X-Ray Back- ph.CO]. ground,” Mon. Not. Roy. Astron. Soc. 370, 213 (2006), [74] T. E. Jeltema and S. Profumo, “Deep XMM Obser- arXiv:astro-ph/0512509 [astro-ph]. vations of Draco Rule out at the 99% Confidence [60] S. Riemer-Sørensen and S. H. Hansen, “Decaying Dark Level a Dark Matter Decay Origin for the 3.5 keV Matter in Draco,” Astron. Astrophys. 500, L37 (2009), Line,” Mon. Not. Roy. Astron. Soc. 458, 3592 (2016), arXiv:0901.2569 [astro-ph.CO]. arXiv:1512.01239 [astro-ph.HE]. [61] S. Horiuchi, P. J. Humphrey, J. Onorbe, K. N. Abaza- [75] J. Franse et al., “Radial Profile of the 3.55 keV Line out jian, M. Kaplinghat, and S. Garrison-Kimmel, “Ster- to R200 in the Perseus Cluster,” Astrophys. J. 829, 124 ile Neutrino Dark Matter Bounds from Galaxies of (2016), arXiv:1604.01759 [astro-ph.CO]. the Local Group,” Phys. Rev. D89, 025017 (2014), [76] E. Bulbul, M. Markevitch, A. Foster, E. Miller, arXiv:1311.0282 [astro-ph.CO]. M. Bautz, M. Loewenstein, S. W. Randall, and R. K. [62] C. R. Watson, J. F. Beacom, H. Yuksel, and T. P. Smith, “Searching for the 3.5 keV Line in the Stacked Walker, “Direct X-Ray Constraints on Sterile Neutrino Suzaku Observations of Galaxy Clusters,” Astrophys. J. Warm Dark Matter,” Phys. Rev. D74, 033009 (2006), 831, 55 (2016), arXiv:1605.02034 [astro-ph.HE]. arXiv:astro-ph/0605424 [astro-ph]. [77] F. Hofmann, J. S. Sanders, K. Nandra, N. Clerc, and [63] D. Malyshev, A. Neronov, and D. Eckert, “Constraints M. Gaspari, “7.1 keV Sterile Neutrino Constraints from on 3.55 keV Line Emission from Stacked Observations X-Ray Observations of 33 Clusters of Galaxies with of Dwarf Spheroidal Galaxies,” Phys. Rev. D90, 103506 Chandra ACIS,” Astron. Astrophys. 592, A112 (2016), (2014), arXiv:1408.3531 [astro-ph.HE]. arXiv:1606.04091 [astro-ph.CO]. [64] D. Iakubovskyi, E. Bulbul, A. R. Foster, D. Savchenko, [78] F. A. Aharonian et al. (Hitomi Collaboration), “Hit- and V. Sadova, “Testing the Origin of ∼3.55 keV Line omi Constraints on the 3.5 keV Line in the Perseus in Individual Galaxy Clusters Observed with XMM- Galaxy Cluster,” Astrophys. J. 837, L15 (2017), Newton,” (2015), arXiv:1508.05186 [astro-ph.HE]. arXiv:1607.07420 [astro-ph.HE]. [65] M. Loewenstein, A. Kusenko, and P. L. Biermann, [79] N. Cappelluti, E. Bulbul, A. Foster, P. Natarajan, M. C. “New Limits on Sterile Neutrinos from Suzaku Obser- Urry, M. W. Bautz, F. Civano, E. Miller, and R. K. vations of the Ursa Minor Dwarf Spheroidal Galaxy,” Smith, “Searching for the 3.5 keV Line in the Deep Astrophys. J. 700, 426 (2009), arXiv:0812.2710 [astro- Fields with Chandra: The 10 Ms Observations,” As- ph]. trophys. J. 854, 179 (2018), arXiv:1701.07932 [astro- [66] H. Yuksel, J. F. Beacom, and C. R. Watson, “Strong ph.CO]. Upper Limits on Sterile Neutrino Warm Dark Matter,” [80] A. Boyarsky, D. Iakubovskyi, O. Ruchayskiy, and Phys. Rev. Lett. 101, 121301 (2008), arXiv:0706.4084 D. Savchenko, “Surface Brightness Profile of the 3.5 keV [astro-ph]. Line in the Milky Way Halo,” (2018), arXiv:1812.10488 [67] E. Carlson, T. Jeltema, and S. Profumo, “Where Do the [astro-ph.HE]. 3.5 keV Photons Come From? A Morphological Study [81] T. Tamura et al., “An X-Ray Spectroscopic Search for of the Galactic Center and of Perseus,” J. Cosmology Dark Matter and Unidentified Line Signatures in the Astroparticle Phys. 2015, 009 (2015), arXiv:1411.1758 Perseus Cluster with Hitomi,” Publ. Astron. Soc. Japan [astro-ph.HE]. 71 (2019), arXiv:1811.05767 [astro-ph.HE]. [68] S. Riemer-Sørensen, “Constraints on the Presence of a [82] C. Dessert, N. L. Rodd, and B. R. Safdi, “The Dark 3.5 keV Dark Matter Emission Line from Chandra Ob- Matter Interpretation of the 3.5-keV Line is Inconsis- servations of the Galactic Centre,” Astron. Astrophys. tent with Blank-Sky Observations,” Science 367, 1465 590, A71 (2016), arXiv:1405.7943 [astro-ph.CO]. (2020), arXiv:1812.06976 [astro-ph.CO]. [69] T. E. Jeltema and S. Profumo, “Discovery of a 3.5 [83] F. Hofmann and C. Wegg, “7.1 keV Sterile Neutrino keV Line in the Galactic Centre and a Critical Look Dark Matter Constraints from a Deep Chandra X-Ray at the Origin of the Line Across Astronomical Tar- Observation of the Galactic Bulge Limiting Window,” gets,” Mon. Not. Roy. Astron. Soc. 450, 2143 (2015), Astron. Astrophys. 625, L7 (2019), arXiv:1905.00916 arXiv:1408.1699 [astro-ph.HE]. [astro-ph.HE]. [70] M. E. Anderson, E. Churazov, and J. N. Bregman, [84] K. N. Abazajian, “Resonantly Produced 7 keV Ster- “Non-Detection of X-Ray Emission from Sterile Neutri- ile Neutrino Dark Matter Models and the Properties of nos in Stacked Galaxy Spectra,” Mon. Not. Roy. Astron. Milky Way Satellites,” Phys. Rev. Lett. 112, 161303 15

(2014), arXiv:1403.0954 [astro-ph.CO]. [101] A. G. A. Brown et al. (Gaia Collaboration), “Gaia [85] D. P. Finkbeiner and N. Weiner, “X-Ray Line from Ex- Data Release 2. Summary of the Contents and Sur- citing Dark Matter,” Phys. Rev. D94, 083002 (2016), vey Properties,” Astron. Astrophys. 616, A1 (2018), arXiv:1402.6671 [hep-ph]. arXiv:1804.09365 [astro-ph.GA]. [86] T. Higaki, K. S. Jeong, and F. Takahashi, “The 7 keV [102] W. Voges et al., “The ROSAT All-Sky Survey Bright Axion Dark Matter and the X-Ray Line Signal,” Phys. Source Catalogue,” Astron. Astrophys. 349, 389 (1999), Lett. B733, 25 (2014), arXiv:1402.6965 [hep-ph]. arXiv:astro-ph/9909315 [astro-ph]. [87] V. Brdar, J. Kopp, J. Liu, and X.-P. Wang, “X-Ray [103] H. Mori, Y. Maeda, Y. Ueda, T. Dotani, and M. Ishida, Lines from Dark Matter Annihilation at the keV Scale,” “Suzaku Observations of Unidentified X-Ray Sources Phys. Rev. Lett. 120, 061301 (2018), arXiv:1710.02146 Toward the Galactic Bulge,” Publ. Astron. Soc. Japan [hep-ph]. 64 (2012). [88] M. H. Namjoo, T. R. Slatyer, and C.-L. Wu, “En- [104] H. An, K. K. Madsen, N. J. Westergaard, S. E. Boggs, hanced N-Body Annihilation of Dark Matter and its In- F. E. Christensen, W. W. Craig, C. J. Hailey, F. A. direct Signatures,” J. High Energy Phys. 03, 077 (2019), Harrison, D. K. Stern, and W. W. Zhang, “In-Flight arXiv:1810.09455 [astro-ph.CO]. PSF Calibration of the NuSTAR Hard X-Ray Optics,” [89] K. Nakayama, F. Takahashi, and T. T. Yanagida, “Re- Proc. SPIE Int. Soc. Opt. Eng. 9144, 91441Q (2014), visiting the Number-Theory Dark Matter Scenario and arXiv:1406.7419 [astro-ph.IM]. the Weak Gravity Conjecture,” Phys. Lett. B790, 218 [105] D. E. Gruber, J. L. Matteson, L. E. Peterson, and (2019), arXiv:1811.01755 [hep-ph]. G. V. Jung, “The Spectrum of Diffuse Cosmic Hard X- [90] L. Gu, J. Kaastra, A. J. J. Raassen, P. D. Mullen, R. S. Rays Measured with HEAO-1,” Astrophys. J. 520, 124 Cumbee, D. Lyons, and P. C. Stancil, “A Novel Sce- (1999), arXiv:astro-ph/9903492 [astro-ph]. nario for the Possible X-Ray Line Feature at ∼3.5 keV: [106] E. Churazov et al., “INTEGRAL Observations of the Charge Exchange with Bare Sulfur Ions,” Astron. Astro- Cosmic X-Ray Background in the 5-100 keV Range via phys. 584, L11 (2015), arXiv:1511.06557 [astro-ph.HE]. Occultation by the Earth,” Astron. Astrophys. 467, 529 [91] L. Gu, J. Mao, J. de Plaa, A. J. J. Raassen, (2006), arXiv:astro-ph/0608250 [astro-ph]. C. Shah, and J. S. Kaastra, “Charge Exchange in [107] H. Kaneda, K. Makishima, S. Yamauchi, K. Koyama, Galaxy Clusters,” Astron. Astrophys. 611, A26 (2018), K. Matsuzaki, and N. Y. Yamasaki, “Complex Spectra arXiv:1710.04784 [astro-ph.HE]. of the Galactic Ridge X-Rays Observed with ASCA,” [92] E. G. Speckhard, K. C. Y. Ng, J. F. Beacom, and Astrophys. J. 491, 638 (1997). R. Laha, “Dark Matter Velocity Spectroscopy,” Phys. [108] T. Yuasa, K. Makishima, and K. Nakazawa, Rev. Lett. 116, 031301 (2016), arXiv:1507.04744 [astro- “Broadband Spectral Analysis of the Galactic Ridge ph.CO]. X-Ray Emission,” Astrophys. J. 753, 129 (2012), [93] D. Powell, R. Laha, K. C. Y. Ng, and T. Abel, “Doppler arXiv:1205.1574 [astro-ph.GA]. Effect on Indirect Detection of Dark Matter Using Dark [109] K. Perez, R. Krivonos, and D. R. Wik, “The Matter Only Simulations,” Phys. Rev. D95, 063012 Galactic Bulge Diffuse Emission in Broad-Band X- (2017), arXiv:1611.02714 [astro-ph.CO]. Rays with NuSTAR,” Astrophys. J. 884, 153 (2019), [94] R. A. Krivonos, S. S. Tsygankov, I. A. Mereminskiy, arXiv:1909.05916 [astro-ph.HE]. A. A. Lutovinov, S. Yu. Sazonov, and R. A. Sunyaev, [110] R. K. Smith, N. S. Brickhouse, D. A. Liedahl, and “New Hard X-Ray Sources Discovered in the Ongoing J. C. Raymond, “Collisional Plasma Models with INTEGRAL Galactic Plane Survey after 14 Years of APEC/APED: Emission-Line Diagnostics of Hydrogen-like Observations,” Mon. Not. Roy. Astron. Soc. 470, 512 and Helium-like Ions,” Astrophys. J. 556, L91 (2001), (2017), arXiv:1704.03364 [astro-ph.HE]. arXiv:astro-ph/0106478 [astro-ph]. [95] R. Launhardt, R. Zylka, and P. G. Mezger, “The [111] S. Sazonov, E. Churazov, R. Sunyaev, and M. Revnivt- Nuclear Bulge of the Galaxy. III. Large Scale Phys- sev, “Hard X-Ray Emission of the Earth’s Atmosphere: ical Characteristics of Stars and Interstellar Mat- Monte Carlo Simulations,” Mon. Not. Roy. Astron. Soc. ter,” Astron. Astrophys. 384, 112 (2002), arXiv:astro- 377, 1726 (2007), arXiv:astro-ph/0608253. ph/0201294 [astro-ph]. [112] E. Churazov, S. Sazonov, R. Sunyaev, and M. Revnivt- [96] M. Revnivtsev, S. Molkov, and S. Sazonov, “Map of sev, “Earth X-Ray Albedo for Cosmic X-Ray Back- the Galaxy in the 6.7-keV Emission Line,” Mon. Not. ground Radiation in the 1–1000 keV Band,” Mon. Roy. Astron. Soc. Lett. 373, L11 (2006), arXiv:astro- Not. Roy. Astron. Soc. 385, 719 (2008), arXiv:astro- ph/0605693 [astro-ph]. ph/0608252. [97] F. A. Harrison et al., “The Nuclear Spectroscopic Tele- [113] V. Fioretti, A. Bulgarelli, G. Malaguti, V. Bianchin, scope Array (NuSTAR) High-Energy X-Ray Mission,” M. Trifoglio, and F. Gianotti, “The Low Earth Orbit Astrophys. J. 770, 103 (2013), arXiv:1301.7307 [astro- Radiation Environment and its Impact on the Prompt ph.IM]. Background of Hard X-Ray Focusing Telescopes,” in [98] D. R. Wik et al., “NuSTAR Observations of the Bullet High Energy, Optical, and Infrared Detectors for As- Cluster: Constraints on Inverse Compton Emission,” tronomy V , Vol. 8453, edited by Andrew D. Holland Astrophys. J. 792, 48 (2014), arXiv:1403.2722 [astro- and James W. Beletic, International Society for Optics ph.HE]. and Photonics (SPIE, 2012) p. 833. [99] K. K. Madsen et al., “Calibration of the NuSTAR [114] F. A. Harrison et al. (NuSTAR Collaboration), “The High Energy Focusing X-Ray Telescope,” Astrophys. J. NuSTAR Extragalactic Surveys: The Number Counts Suppl. 220, 8 (2015), arXiv:1504.01672 [astro-ph.IM]. of Active Galactic Nuclei and the Resolved Fraction of [100] T. Prusti et al., “The Gaia Mission,” Astron. Astrophys. the Cosmic X-Ray Background,” Astrophys. J. 831, 185 595, A1 (2016), arXiv:1609.04153 [astro-ph.IM]. (2016), arXiv:1511.04183 [astro-ph.HE]. 16

[115] M. Revnivtsev, S. Sazonov, E. Churazov, W. Forman, timate for Sgr A* from a Multistar Orbit Analysis,” A. Vikhlinin, and R. Sunyaev, “Discrete Sources as the Astrophys. J. 830, 17 (2016), arXiv:1607.05726 [astro- Origin of the Galactic X-Ray Ridge Emission,” Nature ph.GA]. 458, 1142 (2009), arXiv:0904.4649 [astro-ph.GA]. [125] R. Abuter et al. (GRAVITY Collaboration), “A Geo- [116] R. Krivonos, S. Tsygankov, M. Revnivtsev, S. Sazonov, metric Distance Measurement to the Galactic Center E. Churazov, and R Sunyaev, “INTEGRAL Con- Black Hole with 0.3% Uncertainty,” Astron. Astrophys. straints on the Galactic Hard X-Ray Background from 625, L10 (2019), arXiv:1904.05721 [astro-ph.GA]. the Milky Way Anticenter,” Astron. Astrophys. 537, [126] F. Nesti and P. Salucci, “The Dark Matter Halo of A92 (2012), arXiv:1109.2471 [astro-ph.GA]. the Milky Way, AD 2013,” J. Cosmology Astroparticle [117] M. G. Revnivtsev and S. V. Molkov, “Results from a Phys. 1307, 016 (2013), arXiv:1304.5127 [astro-ph.GA]. Deep RXTE/PCA Scan Across the Galactic Plane,” [127] F. Calore, N. Bozorgnia, M. Lovell, G. Bertone, Mon. Not. Roy. Astron. Soc. 424, 2330 (2012). M. Schaller, C. S. Frenk, R. A. Crain, J. Schaye, [118] J. Wilms, A. Allen, and R. McCray, “On the Absorp- T. Theuns, and J. W. Trayford, “Simulated Milky tion of X-Rays in the Interstellar Medium,” Astrophys. Way Analogues: Implications for Dark Matter Indirect J. 542, 914 (2000), arXiv:astro-ph/0008425 [astro-ph]. Searches,” J. Cosmology Astroparticle Phys. 1512, 053 [119] J. M. Dickey and F. J. Lockman, “Hi in the Galaxy,” (2015), arXiv:1509.02164 [astro-ph.GA]. Ann. Rev. Astron. Astrophys. 28, 215 (1990). [128] M. Schaller et al., “Dark Matter Annihilation Ra- [120] P. M. W. Kalberla, W. B. Burton, D. Hartmann, E. M. diation in Hydrodynamic Simulations of Milky Way Arnal, E. Bajaja, R. Morras, and W. G. L. P¨oppel, Haloes,” Mon. Not. Roy. Astron. Soc. 455, 4442 (2016), “The Leiden/Argentine/Bonn (LAB) Survey of Galac- arXiv:1509.02166 [astro-ph.CO]. tic Hi: Final Data Release of the Combined LDS and [129] A. Burkert, “The Structure of Dark Matter Halos in IAR Surveys with Improved Stray-Radiation Correc- Dwarf Galaxies,” Astrophys. J. Lett. 447, L25 (1995), tions,” Astron. Astrophys. 440, 775 (2005), arXiv:astro- arXiv:astro-ph/9504041 [astro-ph]. ph/0504140 [astro-ph]. [130] H.-N. Lin and X. Li, “The Dark Matter Profiles in the [121] J. F. Navarro, C. S. Frenk, and S. D. M. White, “A Milky Way,” Mon. Not. Roy. Astron. Soc. 487, 5679 Universal Density Profile from Hierarchical Clustering,” (2019), arXiv:1906.08419 [astro-ph.GA]. Astrophys. J. 490, 493 (1997), arXiv:astro-ph/9611107 [131] E. Jones, T. Oliphant, P. Peterson, et al.,“scipy: [astro-ph]. Open source scientific tools for python,” https://www. [122] D. Hooper, “The Density of Dark Matter in the Galactic scipy.org/ (2001–), [Online, accessed August 15, 2021]. Bulge and Implications for Indirect Detection,” Phys. [132] T. P. Robitaille et al. (astropy Collaboration), Dark Univ. 15, 53 (2017), arXiv:1608.00003 [astro- “astropy: A Community python Package for ph.HE]. Astronomy,” Astron. Astrophys. 558, A33 (2013), [123] M. Pato, F. Iocco, and G. Bertone, “Dynamical Con- arXiv:1307.6212 [astro-ph.IM]. straints on the Dark Matter Distribution in the Milky [133] A. M. Price-Whelan et al. (astropy Collaboration), Way,” J. Cosmology Astroparticle Phys. 1512, 001 “The astropy Project: Building an Open-science (2015), arXiv:1504.06324 [astro-ph.GA]. Project and Status of the v2.0 Core Package,” Astro- [124] A. Boehle et al., “An Improved Distance and Mass Es- nomical J. 156, 123 (2018), arXiv:1801.02634 [astro- ph.IM].