X-rays from Clusters of Galaxies

The baryons thermalize to >

Todays Material
How do we know that clusters are massive Virial theorem Lensing -X-ray Hydrostatic equilibrium (but first we will discuss x-ray spectra ) Equation of hydrostatic equil (*) What do x-ray spectra of clusters look like *!P=-!g!"(r) where "(r) is the gravitational potential of the cluster (which is set by the distribution of matter) P is gas pressure and !g is the gas density (!f=(!f/x1, !f/x2!f/xn)

K making clusters strong Xray sources- the potential energy of infall is converted into kinetic energy of the gas. Most of the baryons in a cluster are in the X-ray emitting plasma - only 10-20% are in the galaxies. Clusters of galaxies are self-gravitating accumulations of dark matter which have trapped hot plasma (intracluster medium - ICM) and galaxies. (the galaxies are the least important constituent)

106

The First Detailed Analysis

Rood et al used the King (1969) analytic models of potentials (developed for globular clusters) and the velocity data and surface density of galaxies to infer a very high mass to light ratio of ~230. Since "no" stellar system had M/L>12 dark matter was necessary

Virial Theorem (Kaiser sec 26.3)

The virial theorem states that, for a stable, self-gravitating, spherical distribution of equal mass objects (stars, galaxies, etc), the total kinetic energy of the objects is equal to -1/2 times the total gravitational potential energy. PE~1/2GN2m2/R tot= 1/2GM2tot/R tot (dimensional analysis) If the orbits are random KE=1/2PE (virial theorm) Mtot~2R totv2tot/G

Rood 1972- velocity vs position of galaxies in Coma Surface density of galaxies Paper is worth reading ApJ 175,627

2<T>=-<WTOT> T is the time average of the Kinetic energy and W is the time overage of the potential energy
In other words, the potential energy must equal the kinetic energy, within a factor of two. Consider a system of N particles with mass m and velocity v. kinetic energy of the total system is K.E.(system) = 1/2 m N v2 = 1/2 Mtot v2

Binney, J. and Tremaine, S. "The Virial Equations." 4.3 in Galactic Dynamics. Princeton, NJ: Princeton University Press, pp. 211219, 1987

Virial Theorm Actual Use (Kaiser 26.4.2)

Photometric observations provide the surface brightness #light of a cluster. On the other hand, measurements of the velocity dispersion $2v together with the virial theorem give $2v ~ W/M~ GM/R~ G#mass R #mass is the projected mass density. At a distance D the mass to light ratio (M/L)can be estimated as M/L =#mass/#light = $2v /GD%#light where % is the angular size of the cluster. Applying this technique, Zwicky found that clusters have M/L ~300 in solar units The virial theorem is exact, but requires that the light traces the mass-but will fail if the dark matter has a different profile from the luminous particles.

Mass Estimates
While the virial theorem is fine it depends on knowing the time averaged orbits, the distribution of particles etc etc- a fair amount of systematic errors Would like better techniques Gravitational lensing Use of spatially resolved xray spectra

Light Can Be Bent by Gravity

The more massthe more the light is bent
Cluster of galaxies

We see a distortion of the image

faculty.lsmsa.edu Amount and type of distortion is related to amount and distribution of mass in gravitational lens

Basics of Gravitational Lensing

See Lectures on Gravitational Lensing by Ramesh Narayan Matthias Bartelmann or
http://www.pgss.mcs.cmu.edu/1997/Volume16/ physics/GL/GL-II.html

Lensing
assume the overall geometry of the universe is Friedmann-Robertson- Walker metric matter inhomogeneities which cause the lensing are local perturbations.
Light paths propagating from the source past the lens 3 regimes 1)light travels from the source to a point close to the lens through unperturbed spacetime. 2)near the lens, light is deflected. 3) light again travels through unperturbed spacetime.

mass mass

For a detailed discussion of the problem Rich centrally condensed clusters occasionally produce giant arcs when a background galaxy happens to be aligned with one of the cluster caustics. Every cluster produces weakly distorted images of large numbers of background galaxies. These images are called arclets and the phenomenon is referred to as weak lensing.

The effect of spacetime curvature on the light paths can then be expressed in terms of an effective index of refraction n, which is given by (e.g. Schneider et al. 1992) n = 1 -(2/c2) " As in normal geometrical optics, refractive index n > 1 light travels slower than in free vacuum. effective speed of a ray of light in a gravitational field is v =c/n ~ c-(2/c)"

The deflection of a light ray that passes a point mass M at impact parameter b is

%def =4GM/c2b

Lensing
Due to slower speed of light the signal is delayed by

This is called the Shapiro delay and has been used to obtain the orbits of neutron stars as well

Narayan and Bartellman 1996

 Einstein radius is the scale of lensing For a point mass it is &E =((4GM/c2)(Dds/DdDs))1/2 or in more useful units &E = (0.9")M11 1/2 DGpc-1/2 Lens eq ' = &-(Dds/DdDs)4GM/&c2. or ' =&-&2E /& ' 2 solutions Any source is imaged twice by a point mass lens
Gravitational light deflection preserves surface brightness because of the Liouville theorm

Lensing

Ways of Thinking About Lensing (Kaiser sec 33.5)

This deflection is just twice what Newtonian theory would give for the deflection of a test particle moving at v = c where we can imagine the radiation to be test particles being pulled by a gravitational acceleration. another way to look at this using wave-optics; the inhomogeneity of the mass distribution causes space-time to become curved. The space in an overdense region is positively curved. light rays propagating through the over-density have to propagate a slightly greater distance than they would in the absence of a the density perturbation. Consequently the wave-fronts get retarded slightly in passing through the overdensity and this results in focusing of rays. Another way : The optical properties of a lumpy universe are, in fact, essentially identical to that of a block of glass of inhomogeneous density where the refractive index is n(r) = (1- 2"(r)/c2) with "(r) the Newtonian gravitational potential. In an over-dense region, " is negative, so n is slightly greater than unity. In this picture we think of space as being flat, but that the speed of light is slightly retarded in the over-dense region. All three of the above pictures give identical results

Hoekstra 2008 Texas Conference

Hoekstra 2008 Texas Conference

What we try to measure with X-ray Spectra

From the x-ray spectrum of the gas we can measure a

mean temperature, a redshift, and abundances of the most common elements (heavier than He). With good S/N we can determine whether the spectrum is consistent with a single temperature or is a sum of emission from plasma at different temperatures. Using symmetry assumptions the X-ray surface brightness can be converted to a measure of the ICM density.

What we try to measure II If we can measure the temperature and density at different positions in the cluster then assuming the plasma is in hydrostatic equilibrium we can derive the gravitational potential and hence the amount and distribution of the dark matter. There are two other ways to get the gravitational potential : The galaxies act as test particles moving in the potential so their redshift distribution provides a measure of total mass. The gravitational potential acts as a lens on light from background galaxies.

Why do we care ? Cosmological simulations predict distributions of masses. If we want to use X-ray selected samples of clusters of galaxies to measure cosmological parameters then we must be able to relate the observables (X-ray luminosity and temperature) to the theoretical masses.

Theoretical Tools
Physics of hot plasmas Bremmstrahlung Collisional equilibrium Atomic physics

Physical Processes
Continuum emission Thermal bremsstrahlung, ~exp(-h(/kT) Bound-free (recombination) Two Photon Line Emission (line emission) L() *( (T, abund) (ne2 V) I( ~*( (T, abund) (ne2 l) Line emission dominates cooling at T<107 K Bremmstrahlung at higher temperatures

Cooling Function

Plasma Parameters
Electron number density ne ~ 10-3 cm-3 in the center with density decreasing as ne~r-2 106<T<108 k Mainly H, He, but with heavy elements (O, Fe, ..) Mainly emits X-rays 1042LX < 1045.3 erg/s, most luminous extended X-ray sources in Universe Age ~ 2-10 Gyr Mainly ionized, but not completely e.g. He and H-like ions of the abundant elements (O Fe) exist in thermal equilibrium

Ion fraction for oxygen vs electron temperature

X-ray Spectra of Clusters

Theoretical model of a collisionally ionized plasma kT=4 keV with solar abundances The lines are 'narrow' Notice dynamic range of 104

X-ray Spectra Data

For hot (kT>3x107k) plasmas the spectra are continuum dominated- most of the energy is radiated in the continuum (lines broadened by the detector resolution)

1 keV Plasma
Theoretical model of a collisionally ionized plasma kT=1 keV with solar abundances The lines are 'narrow' Notice dynamic range of 105

Collsionally Ionized Equilibrium Plasma

Observational data for a collisionally ionized plasma kT=1 keV with solar abundances Notice the very large blend of lines near 1 keV- L shell lines of Fe Notice dynamic range of 107 Ratio of model to a 'pure' H/He plasma This plot is designed to show the lines

Strong Temperature Dependence of Spectra

Line emission Bremms (black) Recombin ation (red) 2 photon green

Relevant Time Scales

The equilibration timescales between protons and electrons is t(p,e) ~ 2 x 108 yr at an 'average' location In collisional ionization equilibrium population of ions is directly related to temperature

! (1,2) =

3m1 2" (kT )3 / 2 2 4 8" m2 n2 Z12 Z 2 e ln ,

3/ 2 -1

ln , + ln(bmax / bmin ) # 40 T & ) ne & ! (e, e) # 3 * 105 ) ' 8 $ ' -3 -3 $ yr ( 10 K % ( 10 cm % ! ( p, p ) = m p / me ! (e, e) # 43! (e, e)

! ( p, e) = (m p / me )! (e, e) # 1800! (e, e)

Ion fraction for Fe vs electron temperature

How Did I Know This??

Why do we think that the emission is thermal bremmstrahlung? X-ray spectra are consistent with model X-ray 'image' is also consistent Derived physical parameters 'make sense' Other mechanisms 'do not work' (e.g. spectral form not consistent with black body, synchrotron from a power law: presence of xray spectral lines of identifiable energy argues for collisional process; ratio of line strengths (e.g. He to H-like) is a measure of temperature which agrees with the fit to the continuum)

Mean Free Path for Collisions/ Energy

Mean-free-path +e ~ 20 kpc < 1% of cluster size
T $ ( 23 ' % 8 " & 10 K #
2

ne ' $ % !3 !3 " kpc 10 cm & #

!1

33 / 2 (kT ) 2 $ p " $e = 8 # nee 4 ln !

At T>3x107 K the major form of energy emission is thermal bremmstrahlung continuum *~ 3x10-27 T 1/2 n2 ergs/cm3/sec- how long does it take a parcel of gas to lose its energy? ,~nkT/* )8.5x10 10yrs(n/10-3)-1 T8 1/2 At lower temperatures line emission is important

Why is Gas Hot

To first order if the gas were cooler it would fall to the center of the potential well and heat up If it were hotter it would be a wind and gas would leave cluster Idea is that gas shocks as it 'falls into' the cluster potential well from the IGM Is it 'merger' shocks (e.g. collapsed objects merging) Or in fall (e.g. rain) BOTH

Physical Conditions in the Gas

the elastic collision times for ions and electrons ) in the intracluster gas are much shorter than the time scales for heating or cooling , and the gas can be treated as a fluid. The time required for a sound wave in the intracluster gas to cross a cluster is given by

Ts~6.6x108yr (Tgas/108)1/2(D/Mpc)
(remember that for an ideal gas vsound=!(.P/!g) (P is the pressure, !g is the gas density, .=5/3 is the adiabatic index for a monoatomic ideal gas )

Hydrostatic Equilibrium Kaiser 19.2

Equation of hydrostatic equil

Hydrodynamics
&! + $ % ( !v) = 0 mass conservation (continuity) &t Dv ! + $P + !$" = 0 momentum conservation (Euler) Dt Ds !T = H # L entropy (heating & cooling) Dt !kT P= equation of state m p

!P=-!g!"(r)
where "(r) is the gravitational potential of the cluster (which is set by the distribution of matter) P is the gas pressure !g is the gas density

Add viscosity, thermal conduction, Add magnetic fields (MHD) and cosmic rays Gravitational potential " from DM, gas, galaxies

Deriving the Mass from X-ray Spectra

density and potential are related by Poissons equation

!2"" = 4/!G
and combining this with the equation of hydrostaic equil

For spherical symmetry this reduces to (1/!g) dP/dr=-d"(r)/dr=GM(r)/r2

With a little algebra and the definition of pressure - the total cluster mass can be expressed as

!.(1/!!P)=-!2"" =-4/G!
or, for a spherically symmetric system 1/r2 d/dr (r2/! dP/dr)=-4/!G!

GM(r)=kTg(r)/Gmp)r (dlnT/dr+dln!g/dr)
k is Boltzmans const, is the mean mass of a particle and mH is the mass of a hydrogen atom

Every thing is observable The temperature Tg from the spatially resolved spectrum The density !gfrom the knowledge that the emission is due to bremmstrahlung And the scale size, r, from the conversion of angles to distance

 The emission measure along the line of sight at radius r, EM(r), can be deduced from the X-ray surface brightness, S(%): EM(r) =4 / (1 + z)4 S(%)/0(T, z) ; r = dA(z) % where 0(T, z)is the emissivity in the detector band, taking into account the instrument spectral response, dA(z) is the angular distance at redshift z. The emission measure is linked to the gas density !g by:

Density Profile
a simple model(the ' model) fits the surface brightness well S(r)=S(0)(1/r/a)2) -3'+1/2 cts/cm2/sec/solid angle Is analytically invertible (inverse Abel transform) to the density profile !(r)=!(0)(1/r/a)2) -3'/2 The conversion function from S(0) to !(0) depends on the detector The quantity 'a' is a scale factor- sometimes called the core radius ------------------------------------------------------------------------------ The Abel transform, , is an integral transform used in the analysis of spherically symmetric or axially symmetric functions. The Abel transform of a function f(r) is given by:

EM(r) = #r !2g(R) Rdr/(R2-r2 )

The shape of the surface brightness profile is thus governed by the form of the gas distribution, whereas its normalization depends also on the cluster overall gas content.

f(r)=1/p$r# dF/dy dy/!(y2-r2)

In image analysis tthe reverse Abel transform is used to calculate the emission function given a projection (i.e. a scan or a photograph) of that emission function. In general the integral is not analytic which makes the

Sarazin sec 5
A geometrical interpretation of the Abel transform in two dimensions. An observer (I) looks along a line parallel to the x-axis a distance y above the origin. What the observer sees is the projection (i.e. the integral) of the circularly symmetric function f(r) along the line of sight. The function f(r) is represented in gray in this figure. The observer is assumed to be located infinitely far from the origin so that the limits of integration are #

Surface Brightness Profiles

It has become customary to use a ''' model (Cavaliere and FescoFumiano)' clusters have <'>~2/3
' - r *2 $ %1 + + ( " % & , rc ) " # 2 m p1 gal 0/ but treat as fitting parameter kT ' -r* I X (r ) . %1 + + ( % & , rc )
2 !3 0 +1 / 2

'Two' Types of Surface Brightness Profiles 'Cored' Central Excess Range of core radii and '

2 (r ) =

20

30 / 2

$ " " #

XMM/Newton A1413 Pratt & Arnaud