Transcript harvest09b - NMSU Astronomy

Galaxy Wind IGM Enrichment
from Star Forming Galaxies: 1<z<3
Insights from CDM Simulations
Chris Churchill
New Mexico State University
Daniel Ceverino (HUJ)
Jessica Evans (NMSU)
Glenn Kacprzak (Swinburne)
Anatoly Klypin (NMSU)
Liz Klimek (NMSU)
1.
Gas phase baryonic
structures are observable in
absorption (bright star
forming galaxies + QSO
absorption lines) with equal
visibility at all redshifts,
including z=1 to z=3
2.
Gas flows into and/or out of
galaxies and baryonic halos
are sensitively probed in UV
absorption lines; cold IGM
gas in, heated gas out
3.
Observations indicate that
z~2-3 galaxies with
moderately high star
formation rates are blowing
out significant amounts of
metal enriched gas
(courtesy C. Steidel)
“down the barrel method”
Campaign by Steidel et al of UV (rest-frame) selected z=2-3 galaxies find winds in
virtually all bright galaxies
Composite spectra of z=2.0-2.6 UV
selected galaxies (R=1300)
3 lensed galaxies (z=2.7-3.0; R=6000)
Mhalo ~ 1012 - 1013 M sun
Lbol ~ 1011 - 1012 L sun (r~1-2 kpc)
SFR ~ 10 - 100 Msun/yr (LIRG-ULIRG)
Vc ~ 150 km/s
(Vesc ~ 450 km/s)
(courtesy C. Steidel)
OUTFLOWS expected to be most common in the redshift desert, where star
formation is most active
We directly observe the IGM enrichment process when it is peaking
We directly observe the interplay (fueling by infall and outflow mass loss)
between galaxy evolution and the baryonic environment in the cosmological
context
Gas expelled at z=1-3 could be the “refueling” material for galaxies at the
present epoch
WINDS may (and probably do) play crucial role
- in shaping mass-metallicity relation in galaxies
- explaining difference between galaxy luminosity and mass functions (low end
and/or high end mismatch)
- heating and chemically enriching of the IGM
- termination of star formation (quenching) in low mass galaxies
and old stellar populations in said galaxies (the red and the dead)
“quasar absorption line (QAL) method”
Neutral hydrogen
(rest-frame velocity)
C IV absorption and z=3 galaxies
For N(CIV) > 1013 cm-2, the galaxy / CIV absorber cross correlation function is equal to
the LBG galaxy auto-correlation function, and it increases by a factor
as the
MgofII,1.5-2.0
C IV, OVI
15
-2
Adelberger etal (2003, 2005)
column density is increased to N(CIV) > 10 cm
(rest-frame velocity)
Clear causal connection of “strong” CIV absorbers seen in QSO spectra with galaxies;
I.e., C IV traces metal enriched gas in vicinity (80 kpc proper) of galaxies
O VI absorption and z=3 galaxies
For N(OVI) > 1013.5 cm-2, the OVI absorber temperatures, kinematics, and rate of
Lyman series
incidence are well explained as winds extending to 50 kpc (proper) associated with
(obs wavelength)
LBGs Simcoe etal (2002)
QSO
sightline
To observer
1.
2.
3.
Galaxies form in the cosmic web
They accrete gas, form stars, and deposit energy/metals into IGM
Extended metal enriched “halos” are observed from z=0 to z=4
Arguably, some of the most physical and visual insights are derived from simulations;
- but need detailed galaxy physics AND cosmological setting - very difficult but crucial
4.5 Mpc
Zooming technique! Adaptive Refinement Tree (ART) - increase spatial resolution in
proportion to where all the action is and track processes with low resolution where its not
Example of stellar particles, and hydro gas density, temperature, and metals (20-50 pc)
stars
density cm-3
temp K
Z solar
1000 kpc
z = 2.3
z = 1.3
CDM Hydrodynamic + N-body Adaptive Refinement Tree (ART) in 10 Mpc box
Kravstov etal (1997); Kravstov (1999); Kravstov, Gnede, & Klypin (2004)
- Radiative (UVB) + collisional heating and cooling (atomic+molecular w/ dust as function of
metallicity) using Cloudy grids Haardt & Madau (1996); Ferland etal (1998)
- Star formation physics based upon 1 pc high resolution simulations Kravstov (2003); Ceverino & Klypin (2008)
- Miller-Scalo IMF Miller-Scalo (1979), Type II and Ia SNe yields fzM* Woosely & Weaver (1995)
- Natural gas hydro only, thermal heating drives winds; no velocity kicks, no rolling dice
Simulations are complex, involving a tonne of physics, some of which needs
extensive testing; Presently, observational data of “halos” and “outflows” are
underutilized for constraining galaxy formation physics in cosmological
simulations…
- how to do it (right)?
1. Use “mock” background quasar absorption line methods
“QAL method”
• Place “quasar beam” sightlines through simulation box, generate absorption profiles
• Shoot through target galaxies, can examine different orientations
• Create grid of sightlines to probe line of sight absorption properties spatially
• Study kinematic, equivalent width, column density, and Doppler b distributions
2. Use “mock” starburst galaxy spectra methods
“down the barrel method”
• Synthesize spectrum of central star forming region of star forming galaxy
• Must account for physical extent of nuclear region
• Can examine different viewing angles
• Study kinematics of profiles, etc.
Generate “observed” spectra, analyze as an observer, quantitatively compare
MOCK QUASAR ABSORPTION “PROBING”
- select a galaxy in the box, select orientation for “sky view”, pass line of sight through box
- line of sight (LOS) is given impact parameter and passed through the entire 10 Mpc box
- record the properties of all gas cells probed by the LOS
4.5 Mpc
QSO
400 kpc
resolution ~ 20-50 pc
Z=1.0 (M = 0.8MMW )
Milky way mass at z=0
Examining Properties of Gas in Absorption
D , V, nH , T , Z/Zsun , fH
gas cell
Vlos
QSO
To observer
b
R
V
R = distance of cell center from galaxy center
b = impact parameter (projected R)
1.
Apply the Cloudy models to obtain photoionization + collisional equilibrium ionization
fractions
2.
Determine density of metal ion in cell, obtain optical depth at line of sight velocity
3.
Synthesize “realistic” spectrum; analyze absorption; tie detected absorption to detected
cells
4.
Examine detected cell properties
@ z=3.55 a low SFR “Lyman Break Galaxy”
temperature
QSO LOS GRID
400 x 400 kpc
 = 20 kpc
density
Halo constructed from stellar
feedback winds…
= 10-2 - 10-6 cm-3
T = 105 - 106.3 K
Z = 0.1 - 1
solar
metallicity
Schematic of Velocity Flows
Filaments inflowing parallel to angular momentum vector (face on). Inner 10 kpc, hot
gas outflows perpendicular to the plane, but is overwhelmed by infalling filaments and
is redirected sideways into metal enriched supershells that entrain cool gas
Entrained material
Metals mix into filaments in inner few hundred kpc, but filaments vigorously fuel the galaxy
Animated Movies (rotation of structure about angular momentum vector of galaxy)
QuickTime™ and a
Video decompressor
are needed to see this picture.
Spatial location of CIV inflow
QuickTime™ and a
Video decompressor
are needed to see this picture.
Spatial location of CIV outflow
QuickTime™ and a
Video decompressor
are needed to see this picture.
Spatial location of OVI inflow
QuickTime™ and a
Video decompressor
are needed to see this picture.
Spatial location of OVI outflow
“DOWN THE BARREL” CIV & OVI ABSORPTION
viewed from
“Side A”
viewed from
“Side B”
Analogous to:
Weiner et al (2009)
Steidel, Pettini, & Shapley
• partial covering is ~ 80% (at any given velocity)
• slow rising blue wing (wind signature) not always apparent
• asymmetric for face-on view (along angular momentum vector)
Observed
3 lensed galaxies
z=2.7-3.0
Red is cB58
face on
edge on
Blended doublet
R = 6000
(courtesy C. Steidel)
RADIAL VELOCITY DISTRIBUTION
Absorption line centroids (not maximum velocity extent)
Observations with respect to H nebular emission (stars)
Observed absorption profile mean centroid is -160 km/s
SFR ~ 10 Msun/yr
<vabs> - 115
outflow (nearside)
(Steidel 1997)
outflow (nearside)
STACKING: IMPACT PARAMETER BINNING
how to get spatial information?
In the real world there are a multitude of background sources- they are just not bright!
To increase signal to noise, select impact parameter bins and co-add spectra in the
reference frame of the intervening absorber
down the barrel
D = 8-40 kpc
D = 40-80 kpc
- when you have 100+ fields you can get
some really good numbers per bin!
D
sky view
side view
STACKING: IMPACT PARAMETER BINNING
Stack 1460 galaxies
Keck LRIS-B spectra
SFR ~ 50 - 100 Msun/yr
down the barrel W=1.61A
D = 8-40 kpc W=1.62A
D = 40-80 kpc W=0.91A
Perform similar experiment with simulation
grids…
LRIS-B mock spectra stacked by observed
impact parameter range
SFR ~ 10 Msun/yr
V=0 @ 1549.5
CIV 1548
(singlet)
Blended doublet @
v(sys) = -390 km/s
-500
courtesy C. Steidel
0
+500
OUTFLOW VELOCITY - STAR FORMATION RATE SCALING
V90 = velocity is defined as the 90%
percentile of the gas with outward
radial velocity greater than the escape
velocity of the galaxy
1000
Each data point is a single galaxy
500
The redshift range is z=1-1.5.
Weiner et al (2009)
Directly compared to outflows found in
DEEP2 galaxies
100
V90 ~ SFR0.5
0.1
10
100
Ceverino et al (in preo)
Weiner et al (2009)
In general, the wind velocity scales with SFR in a manner consistent with Mg II winds
OUTFLOW TO INFLOW EVOLUTION
Distribution of Radial Velocity of Absorbing Cells Giving Rise to Detected Absorption
z=3.5
z=3.5
OVI
CIV
Ly
z=1.0
<v> FWHM
115
223
86
357
-27
129
z=1.0
OVI
CIV
Ly
<v> FWHM
-142
225
-132
200
-78
176
Dominated by
filamentary inflow
CONCLUDING REMARKS
Work is still at a very preliminary level….
It is very expensive to run many galaxies to get statistics on the absorption quantities,
which aren’t really published yet!
We are only making qualitative comparisons at this time, though the absorption line
work has constrained the SFR efficiencies from earlier work
It is clear that cold flows are prominent and required to fuel the continuation of star
formation
The scaling of the outflow velocity with SFR qualitatively is promising in its comparison
with observations
The absorption gas method is probably the most promising in that it incorporates the
sensitivity functions of detecting the gas in observed spectra
(post talk material/fodder for Q/A etc)
2-comp sub-DLA
Two main sights for HI
MgII: Plane of sky, 150<v<80 km s-1
Two sights for CIV
absorption- photoionized,
not a single cloud!
Extended sights for OVI
absorption- photo and
collisionally ionization, not
a single cloud!
plane of sky
+18 kpc
MgII 18 kpc behind pos,
0<v<+100 km s-1
EVOLUTION FROM Z=3.5 TO Z=1
density
• baryons continue to fall into galaxy
• local web thins out
• entrained gas from earlier wind extends to 200 kpc, evolution not symmetric about galaxy
EVOLUTION FROM Z=3.5 TO Z=1
• Xray “coronal” conditions within 80 kpc, non uniform (due to filaments)
• too much gas cooled to T=104 K?
• OVI collisional ionization condition present in post shock filaments
temperature
EVOLUTION FROM Z=3.5 TO Z=1
metallicity
• Even though gas is cooling, metals ejected to 200-300 kpc
• At high z, NB filaments enriched by mixing, but haven fallen into galaxy, at low z, Z~10
• metals spread out in more diffuse lower density gas
-2
400 kpc x 400 kpc QSO Grid: Metallicity vs Galactocentric Distance
- Gas cells contributing to objectively detected absorption lines
Inflow
Lower metallicity
Lower column density
Out to 300 kpc
Outflow
Higher metallicity
Higher column density
Out to 200 kpc
Observed QSO absorption line profiles are result of complicated patterns of gas
kinematics, metallicities, and galactocentric distances (metals correlated to kinematics)