The Nature of

Oleg Ruchayskiy on behalf of participants of the proposal 76480 (AO-14) PI: A. Boyarsky

May 11, 2016

Madrid. XMM-Newton Science Workshop 0 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Dark matter in astrophysics and cosmology One of the most firmly established evidence for new physics

Dark matter in astrophysics

Dynamics of stars in galaxies and gas in galaxies . . . Dynamics of galaxies in galaxy clusters . . . Hot X-ray emitting gas in galaxy clusters. . . Gravitational lensing . . .

Madrid. XMM-Newton Science Workshop 1 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Dark matter in cosmology

The is full of structures today . . . but was very homogeneous in the past

At recombination baryon density 5 contrast δbaryon 10 ∼ − Small overdensities grow linear with Without dark matter today

Density contrast 5 2 δbaryon (1 + z)10− 10− z 1100 ∼ ∼ '

Dark matter is the substance that was not coupled to photons and drove the formation of structures in the Universe

Madrid. XMM-Newton Science Workshop 2 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Neutrino dark matter

Cosmic neutrinos Neutrino seems to be a perfect dark matter candidate: neutral, stable, massive, abundantly produced in the early Universe We know how neutrinos interact and we can compute their primordial 3 number density nν = 3 112cm × − To give correct dark matter abundance the sum of neutrino masses, ∑mν , should be ∑mν 11eV ∼

Tremaine-Gunn bound (1979) Such light neutrinos cannot form small galaxies – one would have to put too many of them and violated Pauli exclusion principle Minimal mass for fermion dark matter 300 400eV ∼ − If particles with such mass were weakly interacting (like neutrino) – they would oveclose the Universe (Ω 3!) ∼

Madrid. XMM-Newton Science Workshop 3 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Beyond neutrino DM: weakly interacting particles

Dark matter cannot be both light and weakly interacting

Possible way out?

Make dark matter particles heavy but keep interaction strength GF ∼ Number of such particles is Boltzmann-suppressed nDM nν  Particles should be stable – a symmetry should protect it from decay This class of dark matter candidates is known as Weakly Interacting Massive Particles (or WIMPs)

A typical WIMP-predicting particle physics model: LHC is searching for these particles, but nothing has been found so far

Madrid. XMM-Newton Science Workshop 4 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 No new physics at electroweak scale. Alternatives?

Madrid. XMM-Newton Science Workshop 5 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Beyond neutrino DM: super-weakly interacting particles No traces of new weakly coupled particles – what are the alternatives?

Keep dark matter light but reduce its interaction strength GF  Number of such particles is low, nDM nν , because they were never produced in the early Universe in large amounts The particles can be light (all the way down to Tremaine-Gunn bound) No need to stabilize them – low mass and feeble interaction cross-section can make them cosmologically long-lived Such particles appear in many particle physics models (sterile neutrino, gravitino, axino, . . . )

Madrid. XMM-Newton Science Workshop 6 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Light decaying dark matter

Two-body decay into two massless particles (DM γ + γ or DM γ + ν) → → 1 Monochromatic decay line E = m c2 γ 2 DM The width of the decay line is determined by Doppler broadening:

∆E vvir 4 2 10− 10− Eγ ∼ c ∼ ÷

R R 2 Decay signal ∝ ρDM Annihilation signal ∝ ρDM Madrid. XMM-Newton Science Workshop 7 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Searches for radiatively decaying dark matter

Φ®ΓΓ 29 10 HEAO-1 INTEGRAL COMPTEL 28 28 10 EGRET 10 FERMI D [sec] τ

27 sec 10 @ 1027

XMM, HEAO-1 SPI Τ Chandra Life-time

1026 τ 8 26 PSD exceeds degenerate Fermi gas = Universe life-time x 10 10

1025 0.01 0.1 1 10 2 3 4 10-1 100 101 102 103 104 10 10 10 MDM [keV] mΦ @MeVD

“Next decade of sterile neutrino studies” [1306.4954] Essig et al.’13

Madrid. XMM-Newton Science Workshop 8 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Two groups reported a line-like feature in X-ray spectra of dark matter-dominated objects

ApJ (2014) [1402.2301]

PRL (2014) [1402.4119]

Energy: 3.5 keV. Statistical error for line position 30 50 eV. ∼ − Lifetime: 1028 sec (uncertainty: factor 0.3 dex) ∼ ∼ Madrid. XMM-Newton Science Workshop 9 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Datasets

Boyarsky, O.R. et al. PRL (2014) [1402.4119]

M31 galaxy 0.98 Msec∆ χ2 = 13.03 .2σ for 2 d.o.f. Perseus off-cluster (MOS) 0.63 Msec∆ χ2 = 9.12 .5σ for 2 d.o.f. Perseus off-cluster (PN) 0.22 Msec∆ χ2 = 8.02 .4σ for 2 d.o.f. Blank sky 15.7 Msec No detection M31 + Perseus (MOS)∆ χ2 = 25.94 .4σ for 3 d.o.f.

Global significance of detecting the same signal in 3 datasets:4 .8σ

Bulbul et al. ApJ (2014) [1402.2301]

73 clusters (XMM, MOS) 6.78 Msec∆ χ2 = 22.84 .3σ for 2 d.o.f 73 clusters (XMM, PN) 2.06 Msec∆ χ2 = 13.93 .3σ for 2 d.o.f 69 distant clusters (XMM, MOS) 4.9 Msec∆ χ2 = 16.53 .7σ for 2 d.o.f. 69 distant clusters (XMM, PN) 4.9 Msec∆ χ2 = 16.53 .7σ for 2 d.o.f...... Perseus center (XMM, MOS) 0.32 Msec∆ χ2 = 12.83 .1σ for 2 d.o.f. Perseus center (Chandra, ACIS-S)∆ χ2 = 11.83 .0σ for 2 d.o.f. Madrid. XMM-Newton Science Workshop 10 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Subsequent works For overview see e.g. [1602.04816] “A White Paper on keV Sterile Neutrino Dark Matter”

Subsequent works confirmed the presence of Perseus the3 .5 keV line in some of the objects T15 Perseus

Boyarsky O.R. et al., Iakubovskyi et al.; Franse et 10-10

θ MOS Clusters al.; Bulbul et al.; Urban et al.; Jeltema & Profumo 2 M31 2 M14 Dwarfs n i challenged it existence in other objects s H14 M31

Malyshev et al.; Anderson et al.; Tamura et al.; 10-11 Sekiya et al. argued astrophysical origin of the line 6.7 6.8 6.9 7.0 7.1 7.2 7.3 7.4 7.5 ms [keV] Gu et al.; Carlson et al.; Jeltema & Profumo; [1507.06655] Riemer-Sørensen; Phillips et al. A common explanation for every detection and non-detection?

When comparing bounds from different objects one should be careful – uncertainty in the dark matter content in each of them results in large

Madrid. XMM-Newton Science Workshop 11 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Dark matter is universal

Dark matter is a universal substance, present in every galaxy (spiral, elliptical, dwarf spheroidal) and every galaxy cluster

From the point of view of astrophysical processes these objects are very distinct!

Madrid. XMM-Newton Science Workshop 12 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Boyarsky, O.R. et al. PRL 115, 161301

1.40 MOS camera 1.30 GC ON, MOS1 GC ON, MOS2 1.20 1.82 Msec 1.10

1.00 4σ+ statistical significance [cts/sec/keV] 0.90 Also detected in 0.80 Normalized count rate S. Riemer-Sorensen 0.70 3.0⋅10-2 [1405.7943]; T. Jeltema &

2.0⋅10-2 S. Profumo [1408.1699]

1.0⋅10-2

0 [cts/sec/keV] 0.0⋅10

-1.0⋅10-2

3.0 3.2 3.4 3.6 3.8 4.0 Energy [keV]

Madrid. XMM-Newton Science Workshop 13 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Non-trivial consistency check Boyarsky, O.R. et al. PRL 115, 161301

GC -1 s -2 10 Perseus

photons cm 27 s M31 -6 s s 27 27 28 s = 2 x 10 τ DM = 6 x 10 = 8 x 10 τ DM τ DM = 1.8 x 10 Line flux, 10 1 τ DM Blank-sky

0.01 0.1 2 Projected mass density, MSun/pc

Madrid. XMM-Newton Science Workshop 14 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Dwarf spheroidal galaxies – smallest DM-dominated objects known

Dwarf spheroidals are “galaxies inside our Galaxy” These are ancient galaxies “swallowed” by the Perfect observational targets: They are “dense” They are “dark” (M/L 102 103) ∼ − They are compact (typical sizes5 0 300) They are nearby (distances 30 100− kpc) − Dwarf spheroidal galaxies are too light and compact to Mass confine X-ray emitting gas (kB T GN ) ∼ Size The best target for XMM – in the of – Draco dSph galaxy

XMM-Newton’s time allocation committee granted us 1.4 Mega-seconds in AO-14 PI: A. Boyarsky Madrid. XMM-Newton Science Workshop 15 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Choice of the target

Geringer-Sameth et al. (2014) Madrid. XMM-Newton Science Workshop 16 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Why XMM?

Signal-to-noise ratio improves as

S Z q = ρDM texp Ωfov A ∆E N · · eff ·

Adopted from XMM is the best among today’s Boyarsky, den Herder et al.’2006; missions, mainly due to its Grasp Neronov, O.R. et al.’2014

Madrid. XMM-Newton Science Workshop 17 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Analysis of Draco dSph Ruchayskiy et al. MNRAS (2016) [1512.07217]

4.5x10-6

4.0x10-6 Distant clusters (MOS)

3.5x10-6 All clusters -6 ]

2 3.0x10 (MOS) The line is detected in the spectrum M31 2.5x10-6 Distant clusters of Draco dSph with low significance 2.0x10-6 (PN) Best-fit Draco PN 2 -6 (∆ = 5 3) Flux [cts/sec/cm 1.5x10 χ . GC 1.0x10-6

Line flux/position are consistent 5.0x10-7

with previous observations: 0.0x100 3.48 3.5 3.52 3.54 3.56 3.58 3.6 E = 3.54 0.06keV, Line position [keV] ± 6 2 F = (1.65 0.7) 10− cts/sec/cm 3.5x10-6 Distant clusters ± × All clusters 3.0x10-6 ] (MOS) There is a shift in position ( 1σ) 2 M31 ∼ 2.5x10-6 between MOS and PN Distant clusters (PN) ∆χ2 = 9 2.0x10-6

The data is consistent with DM ∆χ2 = 4 1.5x10-6 interpretation for lifetime GC ∆χ2 = 1 1.0x10-6 27 PN camera flux [cts/sec/cm τDM > (7 9) 10 sec 5.0x10-7 Best-fit Draco MOS1+MOS2+PN − × 0.0x100 3.46 3.48 3.5 3.52 3.54 3.56 3.58 3.6 3.62 3.64

Line position [keV] Madrid. XMM-Newton Science Workshop 18 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Summary

The nature of dark matter is one of the main puzzles of modern physics We have no idea what is dark matter mass or what type of interactions it has A 3.5 keV spectral feature is consistent with a signal originating from dark matter decays Systematics? – Unlikely: Correctly scales with redshift signal not detected in some long-exposure datasets Astrophysics (e.g. K xviii or S xvi charge exchange)? – Possible, but does not fit all observations! XMM observation of Draco is a critical test! The results from very large AO-14 observational programme support dark matter hypothesis Detecting a line from a dwarf galaxy at 3σ level would be crucial ∼ Requires another observation like this

Madrid. XMM-Newton Science Workshop 19 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Hitomi (Astro-H) spectrum of Perseus

+1.1 5 2 The line in the center of the Perseus cluster has flux2 .1 1.0 10− cts/sec/cm − × (XMM, excluding central1 0 circle) Bulbul et al. ApJ (2014) [1402.2301] The presence of the line is confirmed by Suzaku studies Urban et al. (2014); Franse et al. [1604.01759] Existing 200 ksec of Hitomi spectrum of Perseus galaxy cluster should see this line even with reduce sensitivity!

Information about the line in the Perseus center (its position, intensity, width) will be crucial in deciding on further strategy of searching for 3.5 keV line with XMM and other instruments

Madrid. XMM-Newton Science Workshop 20 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Outlook If this is the dark matte signal. . .

A lot of exciting science Dark matter tomography becomes possible! We will be able to see 3D structures in the Universe — visualized cosmic web histories of major mergers tidal streams 3D structures of galaxy clusters ... X-ray astronomy becomes major cosmological tool of next decades

And a major challenge for particle physics . . . to detect a particle that is 73 times lighter than electron and interacts 10 orders of magnitude weaker than neutrino!

Madrid. XMM-Newton Science Workshop 21 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 XMM-Newton is (and will remain for a long time) the best instrument to discover light decaying dark matter

Thank you for your attention!

Madrid. XMM-Newton Science Workshop 21 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 XMM-Newton is (and will remain for a long time) the best instrument to discover light decaying dark matter

Thank you for your attention!

Madrid. XMM-Newton Science Workshop 21 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Backup slides

Madrid. XMM-Newton Science Workshop 21 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Perseus cluster with Suzaku [1604.01759] Jeroen Franse et al. “Radial Profile of the 3.55 keV line out to R200 in the Perseus Cluster”

Confirms the line’s presence in the Perseus cluster’s core (c.f. Bulbul et al.; Urban et al. in contrast with Tamura et al.) Not sensitive enough to probe off-center surface brightness profile Upper bound consistent with XMM results of (Boyarsky et al.’14)

Madrid. XMM-Newton Science Workshop 21 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Stacked galaxy clusters with Suzaku

Bulbul et al. [1605.02034] 7

TABLE 4 Measured Flux of the 3.5 keV Line in the Stacked Suzaku Clusters

(1) (2) (3) (4) (5) (6) (7) 2 2 proj 2 2 Sample Inst. Energy Flux MDM /DL sin (2✓) 6 2 1 10 2 11 (keV) (10 phts cm s )(dof)(dof)(10M /M pc )(10 )

+0.5 +1.3 +1.4 +3.4 Full Sample FI 3.54 1.0 0.5 ( 0.9)1028.1(1068)4.11(1)1.172.71.4 ( 2.3) +0 .2 +0 .2 +0 .4 +5 .9 BI 3.54 0.9 0.7 ( 0.9)1109.9(1077)1.46(1)1.172.51.9 ( 2.5)

Cool-Core FI 3.54 <1.4 1131.7 (1069) 1.68 (1) 1.06 <5.1 Clusters BI 3.54 <2.1 1143.0 (1072) 0.15 (1) 1.06 <6.1

+1.0 +1.9 +2.6 +4.7 Non-Cool Core FI 3.54 2.0 0.7 ( 1.2)1034.7(1075)6.56(1)1.195.31.8 ( 3.1) Clusters BI 3.54 <5.4 1159.9 (1072) 0.51 (1) 1.19 <14.1

2 1 Note: Columns (2) and (3) are the rest energy and flux of the unidentified line in the units of photons cm s at the 68% (90%) confidence level. Column (4) and (5) show the 2 after the line is added to the total model and change in the 2 when an additional Gaussian component is added to the fit; column (6) is the weighted ratio of mass to distance squared of the samples, and column (7) shows the mixing angle limits measured in each sample. Reported constraining limits are at 90% confidence. Energies are held fixed during the model fitting.

Madrid. XMM-Newton Science Workshop 21 / Oleg Ruchayskiy (NBI) 2 =1130.0The nature (1068 of dark dof). matter Adding in an extra Gaussian observations does22 not reveal significant residuals around model to the MOS spectrum at 3.54 keV does not im- 3.54 keV. Indeed, the first fitting attempt (without an prove the fit significantly (2 =1.68) for an additional Gaussian model at 3.54 keV) is an overall good with 2 degrees-of-freedom and results in a non-detection. The of 1041.3 for 1076 dof. Addition of a Gaussian model im- 90% upper limit on the flux of this line at 3.54 keV is proves the fit by 2 of 6.56 for an extra dof. The best- 6 2 1 1.4 10 photons cm s from this spectrum. The up- +1.0 +1.9 6 fit flux of the line becomes 2.0 0.7 ( 1.2) 10 photons per⇥ limit on the flux can be translated to a mixing angle 2 1 ⇥ 11 cm s . The mixing angle corresponding to this flux of 5.1 10 for a given projected dark matter mass per +2.6 +4.7 11 ⇥ 10 2 is 5.3 1.8 ( 3.1) 10 , which is consistent with the full distance squared for the sample (1.06 10 Msun/Mpc ). ⇥ 6 The mixing angle indicated by the⇥ stacked FI observa- sample. The 90% upper limit to its flux is 3.9 10 2 1 ⇥ phts cm s in the FI observations of the non-cool tions of CC clusters is consistent with the Suzaku full- 10 sample and the previous XMM-Newton detections. clusters with a mixing angle of 1.0 10 . Treating the stacked BI observations⇥ of the NCC clus- The overall fit to the stacked BI observations to CC 2 clusters is acceptable with 2 of 1142.85 for 1068 dof. ters, we obtain an acceptable fit ( of 1159.4 with Adding an extra Gaussian line at 3.54 keV does not im- 1071 dof) without an additional Gaussian model at 3.54 prove the fit significantly and results in a non-detection. keV. Adding an extra Gaussian component at 3.54 keV 6 changes the goodness of the fit by 2 of 0.51 (dof = The 90% upper limit to the flux is 2.1 10 photons 2 1 ⇥ 1). The 90% upper limits to the flux of the line is cm s from this spectrum; the upper limit on the 11 2 6 2 1 mixing angle (< 6.1 10 Msun/Mpc ) from this flux 5.4 10 phts cm s , which corresponds to a mixing ⇥ 10 limit is consistent with⇥ the full-sample and FI detections. angle of 1.4 10 for this sample. ⇥ 3.3. Non-Cool Core Clusters 4. SUMMARY We now examine the FI and BI observations of the Stacking X-ray spectra of galaxy clusters at di↵er- NCC clusters. A total of 2.2 Ms good FI and 1.1 Ms good ent provides a sensitive tool to detect weak BI observations are obtained for this sample. The NCC emission features. This method, tested on the XMM- cluster sample contain 46% of the total FI source counts, Newton observations of 73 clusters (Bu14a), resulted in 45% of the total BI source counts of the full sample. the detection of a very weak unidentified spectral line The redshift has a weighted mean value at 0.11 and the at 3.5 keV. In this work, we take a similar approach projected dark matter mass per distance squared is 1.19 and⇠ stack Suzaku FI (XIS0, XIS3) and BI (XIS1) obser- 10 2 ⇥ 10 Msun/Mpc of the NCC subsample. vations of 47 nearby (0.01

Raw/Cleaned FoV [Msec] [arcmin2]

Draco spectra binned by 65 eV, mos1 (black), mos2 (red), pn (green) MOS1 1.43 / 0.97 318.9 MOS2 1.44 / 1.02 573.5 PN 1.29 / 0.65 543.9

PN 65 eV −1 keV −1

−1 0.2 keV −1 0.1

0.1 normalized counts s normalized counts s 0.05 0.2

5×10−30 −1 keV −1

2 3 4 5 6 7 8 9 10 0 Energy (keV)

−5×10−3 normalized counts s 3 3.5 4 4.5 5 5.5 6 6.5 Energy (keV)

Madrid. XMM-Newton Science Workshop 21 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Draco dSph in 2009 vs. 2015

Draco MOS spectra binned by 65 eV, 2009 year (black), 2015 year (red) Draco PN spectra binned by 65 eV, 2009 year (black), 2015 year (red) −1 −1 0.1 keV keV 1 −1 −1 normalized counts s normalized counts s 0.05 0.2 0.5 2

2 4 6 8 10 2 4 6 8 10 Energy (keV) Energy (keV)

Madrid. XMM-Newton Science Workshop 21 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Jeltema & Profumo (2015)-I

Independent analysis of the 2009+2015 Draco observations was done in [1512.01239] No line detected even in PN camera Much stronger upper limits on flux

Black cross: detection of Ruchayskiy et al. Black down arrow:2 σ upper limit from [1512.07217] Ruchayskiy et al. [1512.07217]

Madrid. XMM-Newton Science Workshop 21 / Oleg Ruchayskiy (NBI) The nature of dark matter 22 Jeltema & Profumo (2015)-II

JP15 [1512.07217] is a very succinct paper. No details of data analysis. Trying to reproduce the results we made a number of modifications of our procedure

6 2 2 PN dataset/data processing Flux, 10− ph/cm /s∆ χ 2015 year dataset (26 obs.) used in R15 +0.67 65 eV bin, 2.3-11 keV, NXB+CXB1 .65 0.70 5.3 −+0.74 65 eV bin, 2.3-11 keV, NXB+CXB no OOT corr.1 .57 0.74 4.3 −+0.67 5 eV bin, 2.3-11 keV, NXB+CXB1 .50 0.71 4.4 −+0.71 5 eV bin, 2.5-5 keV, NXB+CXB1 .56 0.76 3.9 −+0.71 5 eV bin, 2.5-5 keV, NXB1 .18 0.70 2.8 − 2009+2015 years dataset (31 obs.) used in JP15 +0.72 65 eV bin, 2.3-11 keV, NXB+CXB1 .47 0.74 4.2 −+0.66 5 eV bin, 2.5-5 keV, NXB1 .04 0.70 2.2 −

Madrid. XMM-Newton Science Workshop 22 / Oleg Ruchayskiy (NBI) The nature of dark matter 22