galaxy cluster gas motions and astro-h: predictions and challenges from simulations

GALAXY CLUSTER GAS MOTIONS AND ASTRO-H: PREDICTIONS AND CHALLENGES FROM SIMULATIONS

John ZuHone, MIT Kavli Institute

withAurora Simionescu (ISAS/JAXA), Eric Miller (MIT), Mark Bautz (MIT), Maxim Markevitch (NASA/GSFC), and Irina Zhuravleva (Stanford)

ASTRO-H MISSION

• To be launched early next year

• Soft X-ray Spectrometer (SXS):

• 3’x3’ FOV

• ~5 eV spectral resolution

• ~1’ spatial PSF

VELOCITY BROADENING

6.4 6.5 6.6

0.1

110

norm

aliz

ed c

ount

s s−

1 keV

−1

Energy (keV)

data and folded model

sekiya 6−Jun−2011 12:36

vturb = 0 km/svturb = 100 km/svturb = 300 km/svturb = 1000 km/s

6.7

w

xyz

jk

e

po

Perseus simulated spectrum (wabs＊bapec)

Astro-H cluster white paper, arXiv:1412.1176

He-like Fe line at E0 ≈ 6.7 keV

Coma

200 kpc

TurbulenceA2319

200 kpc

Sloshing

Astro-H cluster white paper, arXiv:1412.1176

Inertial Range

Dissipation Scale

Injection Scale

P(k) ~ k-α

k

P(k)

Homogeneous, isotropic turbulence

(Zhuravleva et al 2012)

second-order structure function

velocity autocorrelation ⇔ power spectrum

proportionality of 2D and 3D power spectra

VARYING DISSIPATION SCALES

VARYING INJECTION SCALES

MODEL FOR COMATHERMAL PROPERTIESIsothermal β-model: •β = 2/3•nc = 3×10-3 cm-3

•rc = 300 kpc•T = 8 keV

1. Generate density, velocity fields2. Project along z-axis for SB, line shift, width3. Reblock images to 1.5’ resolution (SXS is 3’x3’)

KINEMATIC PROPERTIES •Gaussian random field:•homogeneous, isotropic•power-law spectrum with upper and lower cutoffs

ERRORS• “cosmic variance errors”

• Natural variations in the velocity field

• 100 realizations of the velocity field taken from the power spectrum

• Decrease with more baselines (more pointings spaced out at distance scales we want to resolve)

“strip” “big cross” “small cross”

“fill” “checkerboard”

COSMIC VARIANCE

ERRORS• “measurement errors”

• Statistical errors on line shift and width from Poisson statistics (increase with increasing line width, decrease with increasing exposure)

• Systematic errors on the line shift from gain uncertainty

• For most configurations, we assume ~100 ks per pointing, except “checkerboard”, for which we assume ~55 ks per pointing

• Assume errors are normally distributed and add them to the map

⇒BIAS CORRECTION

VARYING DISSIPATION

SCALES

Curves are essentially indistinguishable

VARYING INJECTION

SCALES

VARYING SPECTRAL

INDEX

Curves are essentially indistinguishable

Kraichnan (MHD)

KolmogorovBurgers (shocks)

VARYING MACH NUMBERS

1-σ confidence limits onpower

spectrumnormalization and injection

scale

fits to structure function and average line

width

WHAT SHOULD WE DO?“big cross” “checkerboard”

~500 ks total exposure (~100 ks per pointing)

~500 ks total exposure (~55 ks per pointing)

SLOSHING CORES

FLASH simulations from ZuHone, Markevitch, & Johnson 2010

Gas motions of M ~ 0.3-0.5, several hundred km/s

• Can we detect these sloshing motions with Astro-H?

• If so, what effect will these motions have on the shift and shape of spectral lines?

• Can we use this spectral analysis to tell us something about microphysics?

Z-PROJECTION

Inviscid Viscous

SB T

σμ

SB T

σμ

X-PROJECTION

Inviscid Viscous

SB T

σμ

SB T

σμ

Y-PROJECTION

Inviscid Viscous

SB T

σμ

SB T

σμ

ASTRO-H POINTINGS AND REGIONS

INVISCID: X-PROJECTION

INVISCID: Y-PROJECTION

INVISCID: Z-PROJECTION

INVISCID: Z-PROJECTION

VISCOUS: X-PROJECTION

VISCOUS: Y-PROJECTION

VISCOUS: Z-PROJECTION

VISCOUS: Z-PROJECTION

DIFFERENT VISCOSITY, SAME SHAPE

DIFFERENT MOTIONS, SAME SHAPE (SOMETIMES)

SYNTHETIC OBSERVATIONS

ZuHone et al 2014, arXiv:1407.1783

SIMX

http://www.youtube.com/watch?v=fUMq6rmNshc

http://yt-project.org

http://hea-www.harvard.edu/simx/

+ PHOX

http://www.mpa-garching.mpg.de/~kdolag/Phox/(Biffi et al 2012, 2013, MNRAS)

event lists: RA, Dec, E Astro-H event files

0.01

0.1

1

coun

ts su−1

d

2 4 6 8

−0.2

0

0.2

resid

uals

Energy (keV)

0 2 8 17 30 48 68 93 122 154 190

SXI SXS

FITTING SPECTRA: Y-AXIS

Model 1:Single-T APEC

model, w/ thermal broadening

only

exposure time = 200 ks

0.01

0.1

0.02

0.05

0.2co

unts

s!1

6.2 6.3 6.4 6.5 6.6 6.7

!0.1

0

0.1

resid

uals

Energy (keV)

Model 1μ = 291 +9 -12 km/skeV-1

FITTING SPECTRA: Y-AXIS

Model 2:Single-T APEC

model, w/ thermal and velocity broadening

exposure time = 200 ks

0.01

0.1

0.02

0.05

0.2

coun

ts s!

1

6.2 6.3 6.4 6.5 6.6 6.7

!0.1

0

0.1

resid

uals

Energy (keV)

Model 2μ = 224 +15 -18 km/sσ = 276 +14 -15 km/s

keV-1

FITTING SPECTRA: Y-AXIS

Two single-TAPEC

models, w/ thermal broadening

only

exposure time = 200 ks

0.01

0.1

0.02

0.05

0.2

coun

ts s!

1

6.2 6.3 6.4 6.5 6.6 6.7

!0.1

0

0.1

resid

uals

Energy (keV)

Model 3μ1 = 377 +13 -10 km/sμ2 = -51 +16 -20 km/s

keV-1

✓ ✓✖ (ish)

What’s wrong with the line shifts?

REASON 1: NEED A BETTER MODEL

REASON 2: PSF SCATTERING

✓ ✓

Correct for PSF Scattering

A2319

200 kpc

A3667

200 kpc

SUMMARY• What motions will Astro-H measure in galaxy clusters?

• Turbulence in Coma:

• Can’t distinguish between dissipation scales, spectral indices (for sane turbulence models)

• Can distinguish between injection scales, but only for certain spatial configurations of pointings(“big cross” or “checkerboard”)

• Gas sloshing:

• Will be able to measure line shifts and widths from sloshing motions

• It will be difficult to distinguished line broadening from sloshing motions vs. turbulent motions—constraints on viscosity are probably limited

• Need to worry about contamination of photons from cluster center—can bias line shifts significantly

• What do we need? Better spatial resolution! (Athena, SMART-X, etc.)