1. X-ray Signatures of Feedback in Intracluster Gas Megan Donahue Michigan State University Collaborators: Mark Voit, Ken Cavagnolo, Steven Robinson, Don Horner (GSFC)

2. X-ray Signatures of Feedback Metals in the ICM The ICM Luminosity - Temperature relationship. ICM entropy profiles and ICM bubbles.

3. Clusters of galaxies Most massive, gravitationally-bound structures in the universe (~10 15 Mo) Hot gas outweighs the stars by a factor ~10 Dark matter outweighs the baryons by a factor ~ 8.

4. ICM enriched on all scales 22 clusters w/Beppo-SAX Consistent Fe abundances Central excess consistent w/production by BCG. DeGrandi et al. 2001; 2004

5. ICM enriched at high redshift Little evolution seen between z=0.3-1.3. Tozzi et al. 2003; Ettori, et al. 2005

6. Abundance Patterns (Tamura et al. 2004) [Si/Fe], [S/Fe] solar [O/Fe] 0.4-0.7 solar in core of cool clusters; solar elsewhere. No trend with T > 2 keV Consistent with Ia and core-collapse supernova yields + star formation rates scaled from the field. (See ASCA results in Baumgartner et al. astro- ph/0309166 for contrary view.) XMM EPIC & RGS spectra, 22 clusters

7. ICM Luminosity - Temperature Relation Gravity-only physics predicts L ~ T 2. Allowing radiation to cool the gas modifies the relation. e.g. Voit & Bryan 2001

8. Galaxy Formation and the ICM Cooling and condensation into stars brings the L-T relation into agreement w/observations. However, cooling alone produces too many stars (Rees & White 1978; Balogh 2001). Star formation contributes metals and energy to the ICM: this feedback alone may regulate star formation in most galaxies.

9. Galaxy Formation and the ICM: Questions Even with feedback from star formation, simulations still predict too much star formation in central dominant galaxies (e.g. Kravtsov 2005). Feedback in “cooling flows”: the ICM in the centers of most X-ray clusters are radiating too brightly to be supported without additional energy input.

10. “Cooling Flows” Clusters with short central cooling times (t c <t H ). Regular, relaxed, luminous X-ray clusters with peaked central X-ray surface brightness. Fairly common, often with central radio sources and H  nebulae in their cores. The radiation losses must be stabilized by feedback.

11. Observed “cooling flow” spectra (XMM) Peterson et al. 2003 FeXVII and other lines from 1 keV gas not present. Two-temperature or “truncated” cooling flow (at ~T/3 - T/2)

12. Cluster gas cools… Temperature gradients from Chandra and XMM observations. Mass cooling rates closer to star formation rates. Gas cooler than ~1 keV not seen or is very faint. The cooling times are still short. What keeps this gas from cooling further?

13. Hydra A z=0.0522 50 kpc Radio Sources & Cluster Cores Can AGN balance radiative cooling in cluster cores? Bubbles in the ICM: (McNamara, Sarazin, Blanton) Heating occurs, but it’s not clear how the AGN compensates for radiative losses. AGN may be the primary culprit in quenching the cooling in cluster cores: but how to tell? Abell 2052 z=0.0353 50 kpc

14. Radio-quiet cluster cores Peres et al. 1998: 23 clusters with cooling rates > 100 solar masses/year 13: emission line nebulae & strong central radio source 2: strong central radio source but no optical line emission (A2029, A3112) 3: emission lines but weak central radio source. (A478, A496, A2142) 5: no emission lines and little or no radio activity. (A644, A1650, A1651, A1689, A2244)

15. Radio-quiet cluster cores Peres et al. 1998: 23 clusters with cooling rates > 100 solar masses/year 13: emission line nebulae & strong central radio source 2: strong central radio source but no optical line emission (A2029, A3112) 3: emission lines but weak central radio source. (A478, A496, A2142) 5: no emission lines and little or no radio activity. (A644, A1650, A1651, A1689, A2244)

16. Chandra Observations Chose 2 symmetric, relaxed clusters without radio sources, A1650 and A2244. ACIS-S observations sufficient to obtain >150,000 counts for radial deprojection of spectra and surface brightness. Temperature and metallicity gradients measured at lower resolution than density gradient.

17. A2244 & A1650 Feedback free? Radio quiet -- upper limits or detections a factor of 30 or more below the others Z = 0.095 and 0.085 KT ~ 5-6 keV

18. What might have been: Fossil radio lobes and/or X-ray cavities suggestive of earlier radio activity. Temperature gradients sufficient to quench cooling via conduction. Very low central entropy values, suggesting that these clusters are on the verge of a heating episode.

19. What is No fossil lobes out to ~100 kpc A1650 A2244 Donahue, et al. 2005

20. What is No temperature gradients: limited, if any, conduction. Donahue, et al. 2005

21. Entropy Specific entropy (Tn -2/3 ) is more closely related to heating and cooling than temperature alone.  S =  (Heat)/T = (3/2) Nk  [ ln(Tn -2/3 )] –Only radiative cooling can reduce entropy –Only heat input (e.g. shocks) can increase entropy –Compression in a gravitational potential changes T but not Tn -2/3 (adiabatic). –S is directly related to t cool (small S(T), short t cool ) –Convective stability condition: dS/dr > 0 Cooling time t c = (14 Gyr) (S/81 keV cm 2 ) 3/2 (T keV ) -1

22. Entropy Gradients Cool cores with feedback evidence show a remarkable consistency in their entropy profiles: S(r) = S 0 + (r/r 1 )  –S 0 ~ 10 keV cm 2 –  ~ 0.9 - 1.3  is about what one expects as a result of structure formation outside the core (but not necessarily inside the core). Almost all have non-zero S 0.

25. Interpretation of profiles Similarity of profiles could be used to argue against episodic heating. No evidence for entropy inversions r > 10 kpc: suggests energy injection can’t just happen at the center. Entropy floors, small entropy inversions, bubbles show current energy injection.

26. Iron Gradients Significant iron gradients, increasing toward the core measured in most of these systems. The presence of a gradient suggests lack of disturbance (e.g. major mergers.) Quasi-stable core gas?

27. What do we see? High central entropy! 35-50 keV cm 2

28. What do we see? T~5-6 keV => t cool > 10 9 years

29. Comparison L R vs. Central Entropy L R vs. Power-Law Slope

30. Significant Iron Gradients

31. What happened? These cluster cores have not yet cooled to low entropy, and will trigger an outburst in the future. OR The AGN in these clusters have a very low duty cycle, requiring enormous energy injection by AGN in the past.

32. Do radio jets heat the ICM? Perseus Cluster & 3C 84Sound Waves in Perseus 10 kpc

33. Dramatic Heating Events MS0735 (McNamara et al.)Hydra A (Nulsen et al.)

34. AGN Heating in Groups Radio-loud groups (circles) tend toward the low-L, high-T side of L-T relation Croston et al. 2004

35. Chandra Entropy Profiles Core entropy profiles very regular Entropy inversions are minor and lie at r < 10 kpc

36. Episodic Heating  t  10 8 yr (K / 10 keV cm 2 ) 3/2 (T / 5 keV) -1 Heating episodes required every ~10 8 yr Central entropy level remains near input level for most of duty cycle Central entropy input cannot greatly exceed 10-20 keV cm 2

37. Entropy Jump Condition K 2  + 0.84 K 1  K  - 0.16 K 1 v2v2 3(4  ) 2/3 v2v2

38. Core Density Structure Core density profile:  (r) ~ 1/r  (r) = r  (r)  ~ 3 x10 -3 g cm -2

39. Zones of AGN Heating for  r ~ const. Luminosity dominated: L ~  r 2 v 3  K ~ 29 keV cm 2 L 45 2/3  3 -4/3 Energy dominated: E ~  r 3 v 2  K ~ 22 keV cm 2 E 59  3 -5/3 r 10 -4/3 Bubble dominated:  E ~ V bub |dP/dr|  r  K ~ 6 keV cm 2 E 59  3 -5/3 (r/r inj ) -0.18 r 10 -4/3

40. Chandra Entropy Profiles Core entropy profiles very regular Entropy inversions are minor and lie at r < 10 kpc

41. Beyond the Core (  r 2 ~ const) Sustained Luminosity: L ~  r 2 v 3 v ~ 1600 km s -1 L 46 1/3 (T/5 keV) -1/3  K / K ~ 0.4 L 46 2/3 (T/5 keV) -5/3

42. Beyond the Core (  r 2 ~ const) Sustained Luminosity: L ~  r 2 v 3 v ~ 1600 km s -1 L 46 1/3 (T/5 keV) -1/3  K / K ~ 0.4 L 46 2/3 (T/5 keV) -5/3 Preserves shape of original K profile!

43. Entropy Profiles Cooling: Breaks self-similarity Its entropy scale determines M-T, L-T relations Feedback: Prevents overcooling What elevates entropy at large radius? Core Profiles: Suggest AGN heating (~10 45 erg s -1, ~10 8 yr) Extended outburst can elevate entire profile

44. AGN heating? Yes! AGN are almost certainly the primary stabilizing mechanisms for cooling cores at z~0.

45. Next Work Complete entropy profile extraction on other radio quiet clusters (almost done). Test idea that cooling rates ~ star formation rates with RGS and Astro E-2 spectra (faint Fe XVII and O VII lines should be present.) Test deprojection assumptions with realistic hydro simulations.

46. Conclusions Recent ICM abundance measurements show enrichment throughout the cluster. ICM abundance evolution since z~1.3 slow; consistent with supernova rates. Central entropies of nearby clusters with short central cooling times are higher in clusters without radio sources. Cooling and star formation explain the ICM L- T relation (at least at high T). AGN are required to complete the story to regulate star formation in cDs and stabilize cool core clusters.