Direct Measurement of a Magnetic Field at z=0.692 Art Wolfe Regina Jorgenson: IOA Tim Robishaw: UCB Carl Heiles:UCB Jason X. Prochaska:UCSC

Evidence for B fields at z ≈ 1 Far IR vs Radio Correlation Dust vs Synchrotron Emission Correlation Independent of z Appleton etal ’04

Evidence for B fields at z ≈ 1: Statistical Evidence for Faraday Rotation in Mg II Absorption Systems at  1.3 (Bernet etal. ‘08) Mg II Control

Redshifted 21 cm Absorption-line Profiles (Arecibo) AO C 286

Redshifted 21 cm Absorption-line Profiles (Arecibo) AO C Jy 1.0 Jy

Hyperfine energy states: B=0

Hyperfine Energy states: B≠0

BB RCP and LCP exitation

BB

Stokes-Parameter Spectra of DLA-3C286 from GBT D I( ) =[I( )-I c (  c ( ) D V( ) =V( )/I c ( )

Parameters from 21 cm Absorption z= ±  =3.75±0.20 km s -1  0 =0.095±0.006 |B los |=83.9±8.8  G completely unexpected: Dynamo theory predicts weaker B fields in the past!

VLBI Continuum maps of 3C 286

Section Cuts Along lines in (a) (a) (b)

Absorption Geometry 200 pc

Damped Ly  Absorption in DLA-3C286

Optical Absorption Lines in DLA-3C286

Optical Absorption Lines in DLA-3C286 Physical Parameters from HIRES spectra v 21 -v opt =3.8  0.2 km s -1  =3.08±0.13 km s -1 [M/H]=-1.3 Dust-to-Gas = 1% Galactic

Optical Absorption Lines in DLA-3C286 Physical Parameters from HIRES spectra v 21 -v opt =3.8±0.2 km s -1  =3.08  0.13 km s -1 [M/H]=-1.3 Dust-to-Gas = 1% Galactic FOS spectrum [C II] 158  m cooling rate < 1% Galaxy cooling rate  SFR < (1/3)  (  SFR ) Milky Way

Upper Limits on Faraday rotation in 3C 286 at = MHz (Gaensler & Ekers ‘08)

Upper Limits on Faraday rotation in 3C 286 at = MHz (Gaensler & Ekers ‘08) RM obs =-0.1±0.1 rad m -2 RM obs =RM DLA +RM MW RM MW =1.4±0.2 rad m -2 RM DLA <1.9 rad m -2 (95% c.l.) Since RM DLA =2.6(N e /10 19 )B los (1+z) -2 the electron fraction x e ≤1.4  (95% c.l.)

Consequences of Strong B Field Since (B 2 plane /8π) >> (  2 /2), B field dominates midplane pressure. Magnetostatic Equilibrium of Gaseous Disk Predicted: (B 2 plane /8  midplane =  G   /2 Observed: (B 2 plane /8  midplane ≥(715)  G  2 /2 Therefore magnetized H I gas cannot be confined by its self-gravity Minimum l.o.s. gas surface Density  min = 490 M  pc -2 required to confine B field

Zeeman Splitting in Molecular Cloud (NGC 2024) OH and H I Zeeman Splitting Results -B los > 100  G - N(H 2 )>10 22 cm -2 But, no H 2 absorption in HST spectra Inferred surface density:  =1450 M  pc -2 Maximum l.o.s B field: B los =87  G

Does B field arise in a molecular cloud ? Absence of Lyman and Werner absorption implies f(H 2 )<7  Perhaps radio photons traverse molecular gas, while optical Photons only go through atomic gas H II H I H2H2 =21 cm =1216Å

Does B field arise in a molecular cloud ? Absence of Lyman and Werner absorption implies f(H 2 )<7  Perhaps radio photons traverse molecular gas, while optical Photons only go through atomic gas But, absence of OH 1612 MHz absorption implies  los < (1/3)  min H II H I H2H2 =21 cm =1216Å

Interpretation(Transient Configuration) B field Enhancement by Merger-induced Shock (F. Shu) If preshock (B 1 )  = 5  G in disk of galaxy Then postshock (B 2 )  = 100  G if u shock = 250 km s -1 Image does not rule out two foreground galaxies

V int Two Gaseous Disks Colliding

V int Two Redshifts Predicted but only one observed

E Gravitational impulse imparted by merging Elliptical Galaxy Radio beam 3 kpc  

WFPC2 Image of 3C 286 (LeBrun etal ‘97; Chen ‘08) PSF-Subtracted smoothed (0.2  ) Diffuse object on top of QSO -Asymmetry  foreground galaxy Suggests sightline to QSO passes within a few kpc of a galaxy

WFPC2 Image of 3C 286 (LeBrun etal ‘97; Chen ‘08) Filament 2.5  SE of QSO AB(F702W)=23.6 mag Is the filament (a) Outer spiral arm at z abs ? (b) A tidal tail at z abs ?

Conclusions 1.First measurement of galactic B field at z >> 0 results in surprise, since B DLA  20  at 6.4 Gyr Look BackTime -Field is average over neutral gas with (a) velocity dispersion  v =3.75 kms -1 and (b) linear scales from 50 to 200 pc -No evidence for strong SFRs usually associated with large fields 2. Consequences -Magnetic Pressure >> self-gravity of H I gas -Magnetic Pressure may be confined by gravity of molecular gas or -B field enhanced by shocks generated by galaxy merger: merger probability p=R merge  t duration =0.006 to Star formation rates in DLAs may be suppressed by strong B fields

Stokes I and V spectra of 3C 286 (N. Kanekar ‘08) 2 hr. integration on GBT Same ‘S Curve’ shape as Wolfe etal (‘08)

Single Component Two Components

B fields at z ~ 2: Faraday Rotation Statistical Evidence Uncertain (but see Kronberg etal 2008) B field estimates require N e which is highly uncertain Oren & Wolfe ‘95

Polar Representation of Spectrum in Complex Visibility Plane

Dynamo Model 1.Conversion of Poloidal to Toroidal Field

Dynamo Model 2.The  effect: Toroidal back to Poloidal and subsequent amplification

Metallicity Distributions

Dust-to-Gas Ratio Distributions

Si II 1526 Equivalent Width Distributions

H I Column Density Distributions

Star Formation in DLAs ? Do DLAs undergo in situ star formation ? If DLAs undergo in situ star formation, how does the comoving SFR density compare to that of LBGs? Or is star formation at high z confined only to compact objects like LBGs? In that case, what is the relationship between LBGs and DLAs? Are DLAs the neutral-gas reservoirs for star formation in LBGs? Connection between Gas and Stars; Kennicutt-Schmidt Law

Cumulative Comoving SFR Density Predicted by the Kennicutt-Schmidt Relation for z~3 N min =2x10 20 N max = N max =1x10 22 N min =2x10 20

Kennicutt-Schmidt Law for Galaxies  = 1.4 Individual Galaxy

How many DLAs in the UDF with z=[2.5,3.5]?

Evolution of Galaxies onto the Red Sequence (Faber etal ‘07) High cool DLAs Low cool DLAs

Distribution of SFR per unit area inferred from CII * Technique CNM Solutions J bkd =(1/2)  J HM Assumptions Features Bimodal Distribution Star formation traced by l c In situ star formation

Bimodality in (u-r,M r ) plane for Nearby Galaxies (Baldry etal ‘04)

U-R Bimodality in Color

Mass per unitComoving Volume versusredshiftMass per unitComoving Volume versusredshift Current Visible Matter Comoving Density of Neutral Gas Versus Redshift

Mass per unitComoving Volume versusredshiftMass per unitComoving Volume versusredshift Current Visible Matter DLAs are Potential Neutral Gas Reservoirs for Star Formation at High Redshifts

Why is efficiency of in situ star formation low in DLAs? 1.Low Molecular Content of Gas in DLAs --DLA median f H2 = Galaxy median f H2 = DLA Gas is Subcritical --Critical Surface Density for Toomre Instability  crit =  G --Epicyclic Frequency,   H(z), increases with z 3. B fields stabilize gas against collapse

Background Heating and 158  m Emission rates versus gas density (z=3)

DLAs with l c ≤ erg s -1 H -1 observed But for half the population, -- l c obs > l c predicted for heating by Haardt- Madau backgrounds --local heat source required observed --grain photoelectric heating from FUV radiation emitted by massive stars

Threshold SFR/Area versus smoothing kernel diameter Comparison with UDF In situ star formation in ‘high cool’ DLAs ruled out for all  kern In situ star formation in ‘low cool’ DLAs ruled out only for  kern > 0.7” at z=2.5  kern > 1.3” at z=3.0  kern > 2.0” at z=3.5

Physical Interpretation of two DLA populations 2. DLAs with l c ergs s -1 H -1 CNM gas heated by FUV Radiation Emitted by LBGs embedded in DLAs >

[C II] 158  m Cooling Site LBG FUV Photon

Bivariate Distribution of DLAs in ([M/H],W 1526 ) Plane

Bivariate Distribution of DLAs in (  SFR,W 1526 ) plane

Fraction of DLAs with M DM > M  can be large (Nagamine etal ‘07) Origin of Bimodality (cont.) Presence of hot gas (T>3x10 5 K) inferred from OVI absorption in > 20 % of DLAs (Fox etal ‘07)

Test: Metallicity Distribution of DLAs with OVI should Resemble that of high l c DLAs Results: inconclusive, but consistent with bimodality

Summary: K-S Law at high z Neutral Gas with N(HI)=2  to10 22 cm -2 at z=[2.5,3.5] : --Area Covering Factor = 1/3 --Low incidence of extended faint galaxies in UDF implies comoving SFR density at least a factor of 30 lower than predicted by applying K-S Relation to DLAs --Star formation efficiency in DLA gas lower than in nearby galaxies --SFRs may be suppressed by (a) low molecular content and/or (b) high N crit of DLA gas Star Formation at High z: --Area covering factor = 10 -3

Summary: Bimodality in DLAs Bimodality in Damped Ly  Systems: --Evidence for two peaks in l c distribution separated by trough at l c crit --Support from disjoint  v 90, [M/H], [D/G], W 1526 distributions Interpretation: --DLAs with l c  l c crit : Absorption in CNM heated by low level of Star Formation (in situ) (or background heated WNM) --DLAs with l c >l c crit : Absorption in CNM gas in heated by central ‘bulge’ sources (LBGs) (not in situ). Galaxy Formation Context: --Origin of current bimodality in galaxies? Red galaxy sequence: l c >l c crit DLAs: Hot Inflows in M > M  halos Blue galaxy sequence:l c  l c crit DLAs: Cold Inflows in M ≤ M  halos Future Tests: --Reality of [M/H] versus W 1526 correlation --Is l c correlated with incidence of OVI absorption?

(1) Lower SFR efficiency due to low Molecular Content of DLAs Toomre instability produces gravitationally bound clouds:N  >N  crit But clouds cannot cool below ≈ 50 K due to low molecular content of DLA gas - DLA Median f H2 = Galaxy Median f H2 =10 -1 In most models Thus, gravitationally bound atomic clouds do not collapse to form stars

Positive Detections Lower Limits Upper Limits Everything Comparison between Positive Detections and Limits

Comparison between empirical And predicted unimodal l c distributions unimodal

H  and Surface Density Profiles in 2 Spirals (Martin ‘06)

i  logN=21.2 cm -2  d  * /dt=      kpc -2 yr -1 Implied surface brightness at z=3:  V =28.4 mag arcsec -2 Measurable in F606W image from Hubble Ultra Deep Field ⊚ K-S Law at high z Surface Brightness

Predicted Comoving SFR Density

Results of UDF Search with F606W Image Unsmoothed Image (  psf =0.09”): -Found 11,000 objects with V<30.5 mag -None satisfied criteria for in situ star formation at Kennicutt rate: i.e.,  V > 26 mag arcsec -2,  dla > 0.25” Smoothed Images: -Removed high surface-brightness objects:  V < 26 mag arcsec -2 -Smoothed image with Gaussian kernels with FWHM=  kern to enhance SNR when  kern =  dla -Let  kern =0.25” to 4.0” or d dla =1.9 kpc to 31 kpc

Significance of Upper Limits in UDF Z=2.5 Z=3.5  =10 arcmin 2 95 % confidence upper limit, n co < N 95 /  V co Comoving volume  V co =3.2x10 4 Mpc 3 Comoving SFR Density: d  * /dt=n co  ( SFR) Threshold SFR/Area:  d  * /dt) threshold   (I 0 ) threshold

(2) But DLA disks may be sub-critical (Toomre stable) N -3 f(N) f(N) exhibits break at N break = cm -2. If DLAs are randomly oriented disks, N break equals maximum N . Infer N  versus r from f(N) N N -2

(2) But DLA disks may be sub-critical (Toomre stable) N -3 f(N) f(N) exhibits break at N break = cm -2. If DLAs are randomly oriented disks, N break equals maximum N . Infer N  versus r from f(N) N N -2 N break

Lower SFR Efficiencies : DLAs are Toomre Stable R > G  /  2 Centrifugally supported R <   /G  Pressure supported` R Unstable  crit =  G RuRu RdRd Epicyclic frequency

Outline DLA surveys for neutral gas at high redshifts - Statistical quantities: f(N,X),  g, etc. - Neutral-Gas reservoirs for star formation at high z? Application of KS law to DLAs -Implications for star-formation efficiency in DLAs DLA investigations of metal absorption lines - Metal content of high-z Universe - Bimodality in DLAs Magnetic Fields in DLAs Conclusions

(3) Critical Surface Density Increases with Redshift Spherical Collapse Model Prediction:  (z,R 200 )=10H(z) Epicyclic frequency :   (R)=R(d  2 /dR)+4  2 Flat Rotation Curve:  (R)=2 1/2  (z,R 200 )(R 200 /R) Critical Surface Density:  crit =  G Neutral Gas is Subcritical

Galex Measurements of SFR Profiles (Boissier etal ‘06)

Parameter versus l c Diagrams

Halo Mass Distribution of Comoving SFR Density Correlation between disk size and M halo  dla ≈(1 " )(M halo /10 11 M  ) 1/3 Kernel angular bandwidth  dla ≈(0.5  1.5)  kern Kernel with  kern =  2 " sensitive to halos with M halo =(  )M  which contribute ≈ 40% of total SFR density Nagamine etal ‘07

Mass per unitComoving Volume versusredshiftMass per unitComoving Volume versusredshift Current Visible Matter Current Neutral gas Dwarf Galaxies Comoving Density of Neutral Gas versus Redshift

Challenges DLA halo mass distribution continuous Lack of emission below 20J bkd Distribution of J predicted for centrally located LBGs But relation between local and bkd. implies <J  J bkd background Background

l c versus n for background heating Empirical Upper Limits on l c 158  m emission rate predicted for gas heated by X-ray and FUV background radiation True l c intersect background curves In low density WNM

Effect of local heat sources on l c versus n curves

Kennicutt-Schmidt Law for Galaxies  = 1.4 Individual Galaxy

Lower SFR Efficiencies: Effect of decreasing slope 

An LBG associated with a DLA (Moller etal ‘02) 8.4 kpc Ly  Emission [O III] Emission

Thresholds for NGC 5236 H  Map

normal discs starburst galaxies log  (HI+H2) M o / pc 2 log  ( SFR ) M o /yr/kpc 2

Solution: Energy Input from LBGs (l c > ) -Comoving Heating Rate from attenuated FUV LBG radiation: H DLA H DLA =(3.0 ±1.5)  ergs s -1 Mpc -3 -Heating rate can balance measured Comoving Cooling rate C DLA C DLA =(2.0±0.5)  ergs s -1 Mpc -3

H DLA -Predicted comoving heating rate: H DLA <2x10 37 erg s -1 Mpc -3 C -[C II] 158  m cooling rate C=(  ±0.5)x   ergs s -1 Mpc -3 -External energy input required -Grain photoelectric heating  d  * /dt Can in situ star formation in DLAs balance cooling in DLAs with l c > ergs s -1 H -1 ?

Bivariate Distribution of DLAs in ([M/H],l c ) Plane

Consquences of upper limit on comoving star formation Density Upper limit: d  * /dt < M yr -1 Mpc -3 Upper limit: d  * /dt < M  yr -1 Mpc -3 -Predicted [M/H] <-2.2 compared to measured [M/H]=-1.4±0.07 -Source of observed metals? 1.Limit on Metal Production

DLA-LBG cross-correlation function and LBG autorcorrelation function (Cooke etal ‘06)

Threshold Determinations from Simulations 0.5” 1.0” 1.Place 10 3 objects with identical exponential brightness profiles, V magnitudes,  DLA, on UDF image 2.Compute recovery fraction as function of V magnitude 3.In principle threshold given by N recover =N 95 4.In practice we used more conservative threshold given by N recover =200

Nature of Reddening in DLAs Murphy etal 2005

Evidence for Dust Depletion and  Enhancement Nuclear Fe Depletion

[C II] 158  m cooling rates versus Dust-to Gas Ratio

Extend N  to r  r break Molecular gas may be located at r  r break Molecular gas may be Toomre unstable Star formation in DLAs may be present, but in regions sequestered away from the neutral gas

Magnitude-Size Relation for LBGs (Bouwens etal 2004)

(1) Cumulative Comoving SFR Density Predicted by Kennicutt-Schmidt Relation for z=3 (2) For Randomly Oriented Disks

with

Kennicutt-Schmidt Law for Galaxies  = 1.4 Individual Galaxy

Evidence for Threshold Surface Densities at z=0

Kennicutt-Schmidt Law for Galaxies  = 1.4 Individual Galaxy

(2) Effects of Low Molecular Content in DLAs Median f H2 =10 -6 in DLAs. By comparison, f H2 =10 -1 in MW SFR~f H2 in most models for star formation Contrast Between DLAs and MW MW: Since N crit ~N shield ~ cm -2, Toomre instability N>N crit leads to significant molecule formation, and to star formation DLAs: If N shield (=10 22 ) > N crit (= ), Toomre instability leads to gravitationally bound atomic clouds, hence no star formation. Reason for high N shield in DLAs is low dust content (  =0.025) and high FUV radiation intensities (G 0 ~4).

Molecular fraction versus total N(H tot ) Molecular content versus total proton column density

Molecular fraction versus color excess (Tumlinson etal ‘02)

Examples of Lyman Break Galaxies

UDF Search with F606W Image Central matches FUV rest-frame wavelength of 1500 Å for z=[2.5,3.5]. FUV emitted mainly by massive stars, so L (t) proportional to SFR(t) Same technique used to get SFRs for LBGs No U-band sensitivity in UDF. Therefore photometric z’s unreliable. But technique valuable for obtaining upper limits on comoving SFR densities

SFR or Luminosity per unit Comoving volume unit Comoving volume Observed De-reddened Consequences Known star formation occurs in compact objects SFRs higher at large redshifts 50 % of current stars and metals produced by z~1 10 % of current stars and metals produced by z~3

But This Picture is Incomplete At high redshifts most baryons were gas Since then cold, neutral gas condensed into stars and C and heavier elements formed. How did this happen? Throughout gas? SFRs of LBGs imply  gas =  visible (z=0) consumed between z=5 and z=2 External Neutral gas reservoirs may be needed

DLAS are Definition of Damped Ly  System (DLA): N(HI)≥ 2  cm -2 Distinguishing characteristics of DLAs : (1) Gas is Neutral (2) Metallicity is low: [M/H]=-1.3 (3) Molecular fraction is low:f H2 ~10 -5 Si II Fe II Optical : Damped Ly  Absorption

Metal Weak Metal Strong

Slide 17r NNNN N0N0N0N0

Observed H I Column-Density Distribution Function

Threshold Determinations from Simulations 0.5” 1.0” 2.0” 4.0”  DLA 1.Place 10 3 objects with identical V magnitude,  DLA, on UDF image 2.Compute recovery fraction as function of V magnitude 3.In principle threshold given by N recover =N 95 4.In practice we used more conservative threshold given by N recover =200

Why is Star Formation Less Efficient at High z ? 1.Critical Surface Density Increases with redshift  crit (R disk )=(2 1/2 V c /R disk )  /  G -Spherical Collapse Model: V c /R 200 =10H(z) -Spherical Collapse Model: V c /R 200 =10H(z)

High z threshold

2. Does Low Molecular Content of DLAs imply low SFR? -But R(HII) in Nearby Galaxies is independent of f GMC independent of f GMC -Implication is that gravitational (Toomre) instability of gas is sufficient condition for star formation of gas is sufficient condition for star formation 3. Lower gas volume densities could suppress SFR

Consequences of upper limit on comoving SFR density Upper limit: d  * /dt < M yr -1 Mpc -3 Upper limit: d  * /dt < M  yr -1 Mpc -3 -Predicted [M/H] < -1.9 compared to measured [M/H]=-1.4±0.07 -Source of Observed Metals? 1.Limit on Metal Production 2.Limit on Energy Input from stars into neutral gas

Are DLAs Passive Layers of Gas? Element Abundance “Floor” implies enrichment process exceeding that of Intergalactic Medium Energy Input into ~50% of Known DLAs exceeds radiative heating rate due to Ultra-Violet and Soft X-ray background radiation --Heating Rates inferred from measured cooling rate per H atom. --Cooling rates measured by inferring [C II] 158  m emission

Does Global Heating Rate Balance Global DLA Cooling Rate? Global Cooling Rate per unit Comoving Volume: C=<l c  gas  crit  m H  =(2±0.5)x10 38 ergs s -1 Mpc -3 Global Heating Rate (grain photoelectric effect): --  is the dust-to-gas ratio --  is the grain photoelectric heating efficiency H DLA <4x10 37 erg s -1 Mpc -3 H LBG =(3±2)x10 38 ergs s -1 Mpc -3 (uncorrected for extinction)

Summary of Results Application of Kennicutt-Schmidt Law to neutral gas at high z predicts significantly higher comoving SFR densities than observed Physical Implications -SFR Efficiency is lower in high-z neutral gas -Explanation ? Increase in critical density with z lower molecular content of gas lower volume density Astrophysical Implications -Predicted metal content lower than observed at z~3 -Comoving cooling rate of gas exceeds upper limits on heating rate due to in situ star formation in gas Suggested Scenario -DLAs with higher [C II] cooling rates powered by centrally located LBGs, which may also supply required metals Star formation mode: central bulge formation at z > 2. Switch to wide spread star formation at lower z’s

Emerging Picture At high z, star formation and element production occurs mainly in compact LBGs. DLAs are neutral gas reservoirs that supply “fuel” for LBG star formation rate. Questions (1) How does DLA gas become chemically enriched? DLA abundance pattern indicates core-collapse SNe. (2) Why is star formation suppressed in DLA gas? (3) Galaxies at z < 1.5 exhibit star formation over large areas. What causes shift from compact to diffuse mode?

Search for LBGs Associated With DLAs Keck R-band Image with 5 U-band dropouts associated With z=2.936 DLA 7h -1 Mpc 10h -1 Mpc  z less than 900 km s -1

Detection of [C II] 158  m Emission from DLAs BenefitsBenefits -Determine spatial extent of cold gas -Determine spatial extent of cold gas -Determine DM mass from line widths -Determine DM mass from line widths -Determine total cooling and heating rates -Determine total cooling and heating rates Is it possible?Is it possible? - -Detected in all late-type galaxies -ALMA required for high-z detection

[C II] contours superposed on 6.75  m Image

[C II] Flux Densities Predicted for DLAs

Predicted S 0 for DLA A 3  Alma limit for 20 hr integration time3  Alma limit for 20 hr integration time 90 % Mass range predicted for CDM Models of DLAs90 % Mass range predicted for CDM Models of DLAs M H I =m D MM H I =m D M

Predicted Distribution of 158  m Flux Densities (Nagamine etal ‘06)

Metal Strong DLA at z=2.6: Probing Nucleosynthesis

Abundance Pattern at z=2.6

FUV Photon Ionizing Photon Grain Grain Photoeletric Heating of Neutral Gas in DLAs H II Region Electron

Identifying Galactic Gas in Absorption Against Quasars Absorption-line Strength Independent of Galaxy Luminosity

Comparison between predicted and observed metal abundance floor