Constraints on Dark Matter Microphysics from Dwarf Galaxies · Ethan Nadler. The Dark Matter Landscape Credit: T. Tait. Buckley & Peter 2018 Microphysical dark matter properties affect

Constraints on Dark Matter Microphysics from Dwarf Galaxies

LIneA Webinar8/6/2020

Ethan Nadler

The Dark Matter Landscape

Credit: T. Tait

Buckley & Peter 2018

Microphysical dark matter properties affect structure formation on small scales

Dark Matter and Structure Formation

Dark Matter and Structure Formation

Planck Collaboration 2019

Planck Collaboration 2019

• Small scales contain information about a variety of dark matter physics: we are compelled to search there!

• What is the luminous content of the smallest halos?

The Galaxy–Halo Connection

Wechsler & Tinker 2018

?

Credit: J. Carlin

MW halo virial radius

Ultra-faint Milky Way satellite galaxies occupy the smallest luminous dark matter halos

Our Census of the Faintest Galaxies

The Milky Way Satellite Population

Credit: K. Bechtol

�60�

�30�

0�

+30�

+60�

Classical

SDSS

PS1

DES

DECam

HSC

ATLAS

GaiaAnt II

Aqr II

Boo IBoo II

Boo III

Boo IV

CVn ICVn II

CarCar II

Car III

Cen I

Cet II

Cet III

Col I

Com

Crt II

Dra

Dra II

Eri II

For

Gru IGru II

Her

Hor I

Hor II

Hyd II

Hyi ILMC

Leo I

Leo II

Leo IVLeo V

Peg III

Phe II

Pic I

Pic II

Psc IIRet II

Ret III

Sgr

Sgr II

Scl

Seg 1

Seg 2

Sex

SMC

Tri II

Tuc II

Tuc IIITuc IV

Tuc V

UMa I

UMa II

UMi

Vir IWil 1

Drlica-Wagner & Bechtol et al. 2020

The Milky Way Satellite Population

Missing Satellites?

“although there seem to be enough observed normal-sized galaxies to match the simulated size distribution, the number of dwarf galaxies is orders of magnitude lower than expected from simulation.” “We show that there is a match between the observed satellite counts

corrected by the detection efficiency of the Sloan Digital Sky Survey … and the number of luminous satellites predicted by CDM, assuming an empirical relation between stellar mass and halo mass. The “missing satellites problem”, cast in terms of number counts, is thus solved.”

“Astronomers couldn’t find enough satellite galaxies orbiting the Milky Way. Now they have the opposite problem, suggesting that our understanding of how galaxies get built is incomplete.”

🤕

https://en.wikipedia.org/wiki/Galaxy

https://en.wikipedia.org/wiki/Dwarf_galaxy

https://en.wikipedia.org/wiki/Orders_of_magnitude

https://en.wikipedia.org/wiki/Orders_of_magnitude

• No missing satellites after correcting for observational completeness and extrapolating a standard stellar mass–halo mass relation

• Consistency has only been demonstrated at fixed modeling assumptions, and only for the brightest half of the observed satellites

No Missing Satellites?

Bullock & Boylan-Kolchin 2017Wetzel et al. 2016

7 7.5 8 8.5 9.0 9.5 10.0log(Mpeak/M�)

0

0.25

0.5

0.75

1.0

f gal

-18-15-12-9-6-30MV [mag]

1

10

50

100

N(<

MV)

-12 -9 -6 -3 0MV [mag]

0

1

2

3

4

5

dN/d

MV

PS1

Observed

Predicted

1. Resimulate Milky Way- like halos from large cosmological volume.

2. Paint satellite galaxies onto subhalos using galaxy—halo model.

4. Calculate likelihood of observed satellites given galaxy—halo connection parameters.

3. Apply observational selection functions based on imaging data.

Mar

kov

Cha

in M

onte

Car

lo↵

<latexit sha1_base64="+wSBPeL8nxBdvzPXA2qswhGhfpg=">AAAB7XicbVBNS8NAEJ3Ur1q/qh69LBbBU0mqoMeiF48V7Ae0oUy2m3btZhN2N0IJ/Q9ePCji1f/jzX/jts1BWx8MPN6bYWZekAiujet+O4W19Y3NreJ2aWd3b/+gfHjU0nGqKGvSWMSqE6BmgkvWNNwI1kkUwygQrB2Mb2d++4kpzWP5YCYJ8yMcSh5yisZKrR6KZIT9csWtunOQVeLlpAI5Gv3yV28Q0zRi0lCBWnc9NzF+hspwKti01Es1S5COcci6lkqMmPaz+bVTcmaVAQljZUsaMld/T2QYaT2JAtsZoRnpZW8m/ud1UxNe+xmXSWqYpItFYSqIicnsdTLgilEjJpYgVdzeSugIFVJjAyrZELzll1dJq1b1Lqq1+8tK/SaPowgncArn4MEV1OEOGtAECo/wDK/w5sTOi/PufCxaC04+cwx/4Hz+AIzPjxw=</latexit>

B<latexit sha1_base64="rCx6YZKyFarkehd6/kmCVau7Ixc=">AAAB8nicbVDLSgMxFM3UV62vqks3wSK4KjNV0GWpG5cV7AOmQ8mkmTY0kwzJHaEM/Qw3LhRx69e482/MtLPQ1gOBwzn3knNPmAhuwHW/ndLG5tb2Tnm3srd/cHhUPT7pGpVqyjpUCaX7ITFMcMk6wEGwfqIZiUPBeuH0Lvd7T0wbruQjzBIWxGQsecQpASv5g5jAhBKRtebDas2tuwvgdeIVpIYKtIfVr8FI0TRmEqggxviem0CQEQ2cCjavDFLDEkKnZMx8SyWJmQmyReQ5vrDKCEdK2ycBL9TfGxmJjZnFoZ3MI5pVLxf/8/wUotsg4zJJgUm6/ChKBQaF8/vxiGtGQcwsIVRzmxXTCdGEgm2pYkvwVk9eJ91G3buqNx6ua81WUUcZnaFzdIk8dIOa6B61UQdRpNAzekVvDjgvzrvzsRwtOcXOKfoD5/MHcyGRXA==</latexit>

�M<latexit sha1_base64="w6kf9ugSeuomJ3tKQ5nBZFNm7MY=">AAAB73icbVDLSgNBEOz1GeMr6tHLYBA8hd0o6DHoxYsQwTwgWcLsZDYZMo91ZlYIS37CiwdFvPo73vwbJ8keNLGgoajqprsrSjgz1ve/vZXVtfWNzcJWcXtnd2+/dHDYNCrVhDaI4kq3I2woZ5I2LLOcthNNsYg4bUWjm6nfeqLaMCUf7DihocADyWJGsHVSu2vYQODeXa9U9iv+DGiZBDkpQ456r/TV7SuSCiot4diYTuAnNsywtoxwOil2U0MTTEZ4QDuOSiyoCbPZvRN06pQ+ipV2JS2aqb8nMiyMGYvIdQpsh2bRm4r/eZ3UxldhxmSSWirJfFGccmQVmj6P+kxTYvnYEUw0c7ciMsQaE+siKroQgsWXl0mzWgnOK9X7i3LtOo+jAMdwAmcQwCXU4Bbq0AACHJ7hFd68R+/Fe/c+5q0rXj5zBH/gff4A8YOP5w==</latexit>

-12 -9 -6 -3 0MV [mag]

0

1

2

3

4

5

dN/d

MV

DES

Observed

Predicted

�gal<latexit sha1_base64="P99ez1bWyq6CFJ3tVJiBTJPQvlE=">AAAB/HicbVBNS8NAEN3Ur1q/oj16CRbBU0mqoMeiF48V7Ae0IWy2m3TpbhJ2J2II8a948aCIV3+IN/+N2zYHbX0w8Hhvhpl5fsKZAtv+Nipr6xubW9Xt2s7u3v6BeXjUU3EqCe2SmMdy4GNFOYtoFxhwOkgkxcLntO9Pb2Z+/4FKxeLoHrKEugKHEQsYwaAlz6yPFAsF9vIR0EfIQ8yLwjMbdtOew1olTkkaqETHM79G45ikgkZAOFZq6NgJuDmWwAinRW2UKppgMsUhHWoaYUGVm8+PL6xTrYytIJa6IrDm6u+JHAulMuHrToFhopa9mfifN0whuHJzFiUp0IgsFgUptyC2ZklYYyYpAZ5pgolk+laLTLDEBHReNR2Cs/zyKum1ms55s3V30Whfl3FU0TE6QWfIQZeojW5RB3URQRl6Rq/ozXgyXox342PRWjHKmTr6A+PzB8RalX4=</latexit>

Simulated LMC

M50<latexit sha1_base64="gV1REod6INZYF0m2I1PcPtZcxYU=">AAAB+XicbVDLSsNAFL2pr1pfUZduBovgqiRV0WXRjRuhgn1AG8JkOm2HTiZhZlIoIX/ixoUibv0Td/6NkzYLbT0wcDjnXu6ZE8ScKe0431ZpbX1jc6u8XdnZ3ds/sA+P2ipKJKEtEvFIdgOsKGeCtjTTnHZjSXEYcNoJJne535lSqVgknvQspl6IR4INGcHaSL5t90OsxwTz9CHz0ysn8+2qU3PmQKvELUgVCjR9+6s/iEgSUqEJx0r1XCfWXoqlZoTTrNJPFI0xmeAR7RkqcEiVl86TZ+jMKAM0jKR5QqO5+nsjxaFSszAwk3lOtezl4n9eL9HDGy9lIk40FWRxaJhwpCOU14AGTFKi+cwQTCQzWREZY4mJNmVVTAnu8pdXSbtecy9q9cfLauO2qKMMJ3AK5+DCNTTgHprQAgJTeIZXeLNS68V6tz4WoyWr2DmGP7A+fwBvspOG</latexit>

EN & Wechsler et al. 2020

• Gaia is revolutionizing our understanding of Milky Way halo properties: mass, concentration, assembly history

• We analyze halos that 1) resemble the MW in these properties, 2) experience a Gaia-Enceladus merger, and 3) have a realistic Large Magellanic Cloud analog

Simulating Milky Way Analogs

Mao, Williamson, Wechsler 2015Callingham et al. 2018

GE-like event

Gaia-Enceladus

Koppelman et al. 2018 Simulated MWs

MW

LMCGaia-Enceladus

Credit: Ralf Kaehler

EN, Y.-Y. Mao, G. Green, R. Wechsler 2019

Empirical modeling allows us to marginalize over theoretical uncertainties

Galaxy–Halo Connection Model

Mock Satellite Observations

simulated “LMC”

simulated “LMC”


DES

Pan-STARRS


simulated “LMC”

Allows for a rigorous statistical comparison to observations

Detection probability is constrained directly by photometric data

-9-6-30MV [mag]

1

5

10

20

N(<

MV)

DES

Observed Satellites

No LMC

Fiducial (with LMC)

-9-6-30MV [mag]

1

5

10

20

N(<

MV)

PS1

Observed Satellites

No LMC

Fiducial (with LMC)

Predicted Satellite Populations• Recent infall (within last 2 Gyr) of LMC system required to fit the data

• Predict 5 (DES) & 1 (PS1) LMC satellites, consistent with Gaia proper motions!


107 108 109 1010 1011

Mpeak [M�]

102

104

106

108

M⇤

[M�]

DES Satellites

PS1 Satellites

Behroozi et al. 2013

Fiducial (with LMC)

�16

�12

�8

�4

0

4

MV

[mag

]

107 108 109 1010 1011

Mvir [M�]

7 7.5 8 8.5 9.0 9.5 10.0log(Mpeak/M�)

0

0.25

0.5

0.75

1.0

f gal

Resolution Limit

Fiducial (with LMC)

The Faint-End Galaxy–Halo Connection

No evidence for a galaxy formation cutoff due to reionization or dark matter physics

atomic cooling limit

observational threshold

-18-15-12-9-6-30MV [mag]

1

10

100

300

N(<

MV)

Without H2 Star Formation

With H2 Star Formation

All Known Satellites

DES + PS1 (weighted)

Fiducial (with LMC)

FIRE Simulations

3 4 5 6 7 8 9

log(M⇤/M�)

The Total MW Satellite Population

• Predict ~220 total MW satellites, 25% of which are LMC-associated

• Vera C. Rubin Observatory should discover ~100 new systems

• Constraints on faint-end slope and galaxy formation efficiency inform hydrodynamic simulations

Munshi et al. 2019

• Spectroscopic follow-up (e.g., on GSMTs) will be key


Interplay with High-z Galaxy Formation

Weisz & Boylan-Kolchin 2017

• Combining the star formation histories of Local Group dwarfs with our abundance matching results will constrain the high-redshift luminosity function

Boylan-Kolchin et al. 2015

Interplay with Local Satellite Surveys

Geha et al. 2017

• Surveys of Milky Way-like satellite systems in the Local Universe are rapidly progressing

• The Satellites Around Galactic Analogs (SAGA) survey targets MW analogs within 20-40 Mpc

• Detailed work to jointly fit the MW and other systems in progress!

• MW satellite luminosity function is typical among SAGA, LV systems

The Faint-End Galaxy–Halo Connection

Wechsler & Tinker 2018

?

CDM WDM

Lovell et al. 2011

Dark Matter Constraints

CDM WDM


free-streaming due to large primordial velocity dispersion → cutoff in abundance of low-mass halos

Bullock & Boylan-Kolchin 2017Pacucci et al. 2013

DM Physics and the Minimum Halo Mass

collisional damping due to DM–SM interactions → cutoff in abundance of low-mass halos


Green et al. 2004 Diemand et al. 2005

even “CDM” has a small-scale cutoff!

linear matter power spectrum (sub)halo mass function

interference due to macroscopic de Broglie wavelength → cutoff in abundance of low-mass halos

Armengaud et al. 2017

Du 2019


EN & Drlica-Wagner et al. 2020

Fit the Milky Way satellite population with subhalo mass function suppression, marginalizing over galaxy–halo connection uncertainties and Milky Way halo properties:

Transfer function

Subhalo mass function

Satellite luminosity function

Current satellite observations are sensitive to ~25% suppression in subhalo abundance relative to CDM

Dark Matter Constraints• Our Milky Way satellite analysis yields the strongest astrophysical constraints

to date on a variety of dark matter properties and particle models

• For thermal relic WDM: mWDM > 6.5 keV (95% confidence), comparable to limits from the Lyman-alpha forest, strong gravitational lensing, and stellar stream perturbations (with largely distinct systematics)

• This is the opposite of searching under a lamppost!


Warm Dark Matter Constraints

• Our analysis excludes nearly all remaining parameter space for resonantly-produced sterile neutrino dark matter

• The sterile neutrino interpretation of the 3.5 keV X-ray line is ruled out at >> 99% confidence

• Interpretation: the dark matter free-streaming length must be smaller than the sizes of the halos that host ultra-faint dwarf galaxies


0.1 1 5 10 30 50k [h Mpc�1]

0.05

0.25

0.5

0.75

1.0

Pco

llis

ional/P

CD

M

mWDM = 0.3 keV

mWDM = 1.2 keV

mWDM = 4.7 keV

�0 = 10�27 cm2

�0 = 10�28 cm2

�0 = 10�29 cm2

mWDM = 0.3 keV

mWDM = 1.2 keV

mWDM = 4.7 keV

�0 = 10�27 cm2

�0 = 10�28 cm2

�0 = 10�29 cm2

1016 1014 1012 1010 108

M [M�]• Collisional damping due to DM–baryon scattering at early times suppresses power on small scales

• Mass of the smallest halo allowed to form corresponds to the size of the horizon when

• Minimum observed halo mass yields analytic cross section limit

EN, V. Gluscevic, K. Boddy, R. Wechsler 2019

Interacting Dark Matter Constraints

Hubble rateMomentum transfer rate



K. Maamari et al. in prep

Preliminary


• Implies that the DM de Broglie wavelength is smaller than the sizes of ultra-faint dwarfs

Fuzzy Dark Matter Constraints• Our analysis robustly rules out

fuzzy dark matter masses below 10-21 eV, with very conservative modeling assumptions

• Can be interpreted as a lower limit on the ultra-light axion mass assuming negligible self and Standard Model couplings


• Dark radiation that decays to DM after BBN (“late-forming DM”) suppresses power on small scales

• MW satellites yields an order-of-magnitude improvement on the transition redshift lower bound:

Late-Forming Dark Matter Constraints

S. Das & EN in prep.

• Cutoff in halo abundance is set by the LFDM transition redshift

Preliminary

0

20

40

60

80

N(>

Vpea

k)

w10

w100

w200

w500

CDM

20 25 30 35 40 45

Vpeak [km s�1]

0

0.5

1.0

NSI

DM/N

CD

M

• Our framework also constrains DM properties that suppress subhalo abundances at late times (e.g. DM self-interactions, decays)

• Self-interactions: Sensitivity to SIDM cross sections of ~1 cm2/g at low velocity scales (stay tuned!)

EN, A. Banerjee, S. Adhikari, Y.-Y. Mao, R. Wechsler 2020

Constraining Late-time DM Physics

1 cm2/g for v < 500 km/s

• Decays: Sensitivity to decaying DM with a lifetime of ~10 Gyr, which has been claimed to alleviate the Hubble tension (stay tuned!)

• Our framework also constrains DM properties that suppress subhalo abundances at late times (e.g. DM self-interactions, decays)

S. Mau et al. in prep

Constraining Late-time DM Physics

Preliminary

Vattis et al. 2019

Outlook

• These constraints will continue to improve with advances in (currently conservative) galaxy–halo modeling and LSST satellite discoveries

• Our analysis informs a variety of DM properties (free-streaming scale, de Broglie wavelength, coupling to the Standard Model) and particle models (sterile neutrinos, generalized WIMPs, ultra-light axions)

• Joint modeling of small-scale structure probes is an important area for future work, starting with satellites, streams, and strong lenses

• The lack of a cutoff in the abundance of observable ultra-faint galaxies down to halo masses of ~108 M☉ yields stringent constraints on the galaxy—halo connection and dark matter physics

Thanks!

Susmita Adhikari, Arka Banerjee, Keith Bechtol, Kimberly Boddy, Subinoy Das, Alex Drlica-Wagner, Vera Gluscevic, Greg Green,

Yao-Yuan Mao, Sidney Mau, Mitch McNanna, Risa Wechsler

DES & Pan-STARRS Survey Selection Functions

• Rigorous satellite search over ~75% of the sky

• Masks for dusty regions, background galaxies, etc.

• 17/18 (DES), 19/31 (PS1) satellites recovered by two search algorithms

• Selection functions from catalog-level searches for simulated satellites


0

1

2

3

log 1

0(r

/pc)

8 < D < 16 kpc

Cet IIDra II

Ret IISeg 1

Tri II

Tuc III

16 < D < 32 kpc

Boo II

Boo III

ComGru II

Seg 2

Tuc IITuc IV

Tuc V

UMa II

Wil 1

32 < D < 64 kpc

�10.0 �7.5 �5.0 �2.5 0.0MV

0

1

2

3

log 1

0(r

/pc)

Aqr IIBoo I

Crt II

Dra

Gru I

Hor I Hor IIPhe II

Pic I

Ret III

Sgr II

Sex

UMa I

UMi

Vir I

64 < D < 128 kpc

�10.0 �7.5 �5.0 �2.5 0.0MV

Boo IVCVn I

CVn IICet III

Col IHer

Leo II

Leo IV

Leo V

Peg III

Psc II

128 < D < 256 kpc

�10.0 �7.5 �5.0 �2.5 0.0MV

Eri II

256 < D < 512 kpc

0.0

0.2

0.4

0.6

0.8

1.0

Det

ecti

onE

�ci

ency

DES

Observational Selection Functions


0

1

2

3

log 1

0(r

/pc)

8 < D < 16 kpc

Cet IIDra II

Ret IISeg 1

Tri II

Tuc III

16 < D < 32 kpc

Boo II

Boo III

ComGru II

Seg 2

Tuc IITuc IV

Tuc V

UMa II

Wil 1

32 < D < 64 kpc

�10.0 �7.5 �5.0 �2.5 0.0MV

0

1

2

3

log 1

0(r

/pc)

Aqr IIBoo I

Crt II

Dra

Gru I

Hor I Hor IIPhe II

Pic I

Ret III

Sgr II

Sex

UMa I

UMi

Vir I

64 < D < 128 kpc

�10.0 �7.5 �5.0 �2.5 0.0MV

Boo IVCVn I

CVn IICet III

Col IHer

Leo II

Leo IV

Leo V

Peg III

Psc II

128 < D < 256 kpc

�10.0 �7.5 �5.0 �2.5 0.0MV

Eri II

256 < D < 512 kpc

0.0

0.2

0.4

0.6

0.8

1.0

Det

ecti

onE

�ci

ency

PS1

Observational Selection Functions


Satellites of the LMC

Predict 5 (1) currently observed LMC-associated satellites in DES (PS1), consistent with Gaia PMs!

Also see Patel et al. 2020!

Kallivayalil et al. 2018


↵ = �1.430+0.020�0.015

0.1

0.2

�M

�M = 0.004+0.093�0.000

7.6

8.0

8.4

M50

M50 = 7.51+0.21�0.00

0.8

1.6

B

B = 0.93+0.59�0.33

0.25

0.50

0.75

�ga

l

�gal = 0.03+0.32�0.00

40

80

A

A = 37+36�25

0.5

1.0

1.5

�lo

gR

�log R = 0.63+0.28�0.30

�1.48�1.44�1.40↵

0.8

1.6

n

0.1 0.2�M

7.6 8.0 8.4M50

0.8 1.6B

0.25 0.50 0.75�gal

40 80A

0.5 1.0 1.5�log R

0.8 1.6n

n = 1.07+0.38�0.57

Faint-end slope consistent with global luminosity function

Luminosity scatter < 0.2 dex

Minimum halo mass corresponding to observed satellites < 3 x108 M☉

Disruption consistent with FIRE sims

Galaxy occupation fraction consistent with step function

Measurement of amplitude, scatter, slope of galaxy-halo size relation



cutoff in galaxy occupation

mass scale of suppression due to dark matter physics


Documents

Constraints on Dark Matter Microphysics from Dwarf Galaxies · Ethan Nadler. The Dark Matter Landscape Credit: T. Tait. Buckley & Peter 2018 Microphysical dark matter properties affect