Herschel observations of FIR emission lines in brightest cluster galaxies
A. C. Edge, J. B. R. Oonk, R. Mittal, S. W. Allen, S. A. Baum, H. Boehringer, J. N. Bregman, M. N. Bremer, F. Combes, C. S. Crawford, M. Donahue, E. Egami, A. C. Fabian, G. J. Ferland, S. L. Hamer, N. A. Hatch, W. Jaffe, R. M. Johnstone, B. R. McNamara, C. P. O'Dea, P. Popesso, A. C. Quillen, P. Salome, C. L. Sarazin, G. M. Voit, R. J. Wilman, M. W. Wise
aa r X i v : . [ a s t r o - ph . C O ] M a y Astronomy&Astrophysicsmanuscript no. 14569rev c (cid:13)
ESO 2018September 19, 2018 L etter to the E ditor Herschel observations of FIR emission lines in brightest clustergalaxies ⋆ A. C. Edge1, J. B. R. Oonk2, R. Mittal3, S. W. Allen4, S. A. Baum3, H. B ¨ohringer5, J. N. Bregman6, M. N. Bremer7,F. Combes8, C. S. Crawford9, M. Donahue10, E. Egami11, A. C. Fabian9, G. J. Ferland12, S. L. Hamer1, N. A.Hatch13, W. Ja ff e2, R. M. Johnstone9, B. R. McNamara14, C. P. O’Dea15, P. Popesso5, A. C. Quillen16, P. Salom´e8,C. L. Sarazin17, G. M. Voit10, R. J. Wilman18, and M. W. Wise19 Institute for Computational Cosmology, Department of Physics, Durham University, Durham, DH1 3LE, UK Leiden Observatory, Leiden University, P.B. 9513, Leiden 2300 RA, The Netherlands Chester F. Carlson Center for Imaging Science, Rochester Institute of Technology, Rochester, NY 14623, USA Kavli Institute for Particle Astrophysics and Cosmology, Stanford University, 452 Lomita Mall, Stanford, CA 94305-4085, USA Max-Planck-Institut f¨ur extraterrestrische Physik, 85748 Garching, Germany University of Michigan, Dept. of Astronomy, Ann Arbor, MI 48109, USA H H Wills Physics Laboratory, Tyndall Avenue, Bristol BS8 1TL, UK Observatoire de Paris, LERMA, CNRS, 61 Av. de l’Observatoire, 75014 Paris, France Institute of Astronomy, Madingley Rd., Cambridge, CB3 0HA, UK Michigan State University, Physics and Astronomy Dept., East Lansing, MI 48824-2320, USA Steward Observatory, University of Arizona, 933 N. Cherry Avenue, Tucson, AZ 85721, USA Department of Physics, University of Kentucky, Lexington KY 40506 USA School of Physics and Astronomy, University of Nottingham, University Park, Nottingham NG7 2RD, UK Department of Physics & Astronomy, University of Waterloo, 200 University Avenue West, Waterloo, Ontario, Canada N2L 3G1 Department of Physics, Rochester Institute of Technology, 84 Lomb Memorial Drive, Rochester, NY 14623-5603, USA Department of Physics and Astronomy, University of Rochester, Rochester, NY 14627, USA Department of Astronomy, University of Virginia, P.O. Box 400325, Charlottesville, VA 22904-4325, USA School of Physics, University of Melbourne, Victoria 3010, Australia ASTRON, Netherlands Institute for Radio Astronomy,P.O. Box 2, 7990 AA Dwingeloo, The NetherlandsReceived 30 March 2010 / Accepted
ABSTRACT
The question of how much gas cools in the cores of clusters of galaxies has been the focus of many, multiwavelength studies in thepast 30 years. In this letter we present the first detections of the strongest atomic cooling lines, [C ii ], [O i ] and [N ii ] in two strongcooling flow clusters, A1068 and A2597, using Herschel
PACS. These spectra indicate that the substantial mass of cold molecular gas( > M ⊙ ) known to be present in these systems is being irradiated by intense UV radiation, most probably from young stars. Theline widths of these FIR lines indicate that they share dynamics similar but not identical to other ionised and molecular gas traced byoptical, near-infrared and CO lines. The relative brightness of the FIR lines compared to CO and FIR luminosity is consistent withother star-forming galaxies indicating that the properties of the molecular gas clouds in cluster cores and the stars they form are notunusual. These results provide additional evidence for a reservoir of cold gas that is fed by the cooling of gas in the cores of the mostcompact clusters and provide important diagnostics of the temperature and density of the dense clouds this gas resides in. Key words.
Galaxies: clusters: intracluster medium, Galaxies: clusters: elliptical and lenticular, cD
1. Introduction
The cooling process at the cores of galaxy clusters is highlycomplex: recent
XMM-Newton and
Chandra observations indi-cate that the cooling rates are reduced by an order of magnitudebelow the simple cooling flow models at temperatures below ∼ × K (Peterson & Fabian 2006). These X-ray observations,when linked with the detection of radio jet inflated bubbles in thecores of many of the strongest cooling flows (see McNamara &Nulsen 2007 for a review), suggest that the strong suppression ⋆ Herschel is an ESA space observatory with science instrumentsprovided by European-led Principal Investigator consortia and with im-portant participation from NASA. of gas cooling is related to energy injection into the intraclustermedium by the action of jets and related AGN activity.The detection of substantial masses of molecular gas in thecores of the most rapidly cooling clusters through CO lines(Edge 2001, Salom´e & Combes 2003) and warm H molecu-lar lines in the NIR and MIR (Ja ff e & Bremer 1997, Egami et al.2006) indicates that not all cooling is suppressed and this cooledgas may provide the fuel for future AGN activity. These tracersof molecular gas appear to correlate with the strength of opticallines from ionised gas (Crawford et al. 1999, Edge 2001) and thedust continuum at MIR and sub-mm wavelengths (O’Dea et al.2008). However, the excitation of these various emission linesand the relative importance of energy input from star formation,
1. C. Edge et al.:
Herschel observations of FIR emission lines in brightest cluster galaxies
AGN, cosmic rays and / or the intracluster medium is poorly con-strained (Ferland et al. 2009).One as yet unexplored diagnostic of the properties of the coldgas are the atomic cooling lines found in the FIR, [C ii ], [O i ] and[N ii ]. The unprecedented sensitivity of Herschel (Pilbratt et al.2010) to FIR line emission o ff ers the opportunity to assess theionisation and density of the colder gas for the first time with the[C ii ] line and two principle [O i ] lines. The authors were awarded140 hours of time in an Open Time Key Project (PI Edge) toinvestigate the FIR line and continuum properties of a sampleof 11 brightest cluster galaxies (BCGs) in well-studied coolingflow clusters selected on the basis of optical emission line andX-ray properties. The full goals of the project are to observeat least five atomic cooling lines for each object that cover arange in density and temperature behaviour and obtain a fullysampled FIR spectral energy distribution for systems where sig-nificant star formation is expected. In this paper we present thePhotodetector Array Camera & Spectrometer (PACS, Poglitschet al. 2010) spectroscopy for the two targets observed in theScience Demonstration Phase (SDP), Abell 1068 ( z = . z = . L ⊙ ) and exhibits somecontribution from an AGN (Crawford et al. 1999, O’Dea et al.2008). On the other hand, A2597 is a relatively weak MIR source(Donahue et al. 2007) with a weak CO detection (Salom´e, priv.comm.) but a powerful radio source (Sarazin et al. 1995). Theimplied FIR luminosity of A2597 of 8.8 × L ⊙ is a factor ofaround 30 below that of A1068 (3.5 × L ⊙ ) and, in addi-tion, the fractional contribution from an AGN in the MIR is alsolower.
2. Observations
We have observed the [C ii ] and [O i ] lines at 157.74 µ m and63.18 µ m for A1068 and A2597 with the PACS spectrometer on Herschel . These are the primary cooling lines of the cold gasat a temperature T <
40 K (Kaufman et al. 1999). In addition forA2597 we observed the [N ii ], [O i ] and [O iii ] lines at 121.90 µ m,145.52 µ m and 88.36 µ m. These lines are used to constrain theexcitation and temperature of this gas. Table 1 gives a summaryof the observations.All spectral line observations were taken in PACS choppedline scan (standard faint line) mode with chopping-nodding.The simple pointed observations mode was used for all obser-vations. The data were reduced following the PACS data reduc-tion guide (PDRG) using the PACS Line Spectroscopy scriptfor Point Source Chop / Nod Mode as presented by the PACSICC team during the
Herschel science demonstration phase dataprocessing workshop at ESAC in december 2009. The reduc-tion was performed within the
Herschel
Interactive ProcessingEnvironment (HIPE) version 2.0.0 (Ott 2010), build RC 3. Wehave processed the data from level 0 (raw channel data) to level2 (calibrated spectra) in a number of steps as outlined in thePDRG. Level 0 to 1.0 processing removes the telescope specificstructures from the data. The slopes of the raw channel data arefitted and removed. The signal is converted from data units tovolts per second. Sky coordinate information is added and badpixels and glitches are removed from data. The data is flatfieldedand flux calibrated by applying the ground based nominal re- sponse function as recommended in the PACS spectroscopy per-formance and calibration (PSPC) document. This ground basedresponse calibration is known to yield overestimated fluxes andfollowing the PSPC we divide our fluxes by 1.3 and 1.1 in theblue and red bands. The accuracy of this flux calibration for thePACS spectrometer, at the time of writing, is about 50 percentwithin a given spectral band (PSPC).During the final stage of the reduction, level 1.0 to 2.0, thedata are spectrally and spatially rebinned into a 5 × × lambdacube. Using the standard 5 × ′′ × . ′′ on the sky. The spec-tral rebinning is performed using the recommended weak linedensity i.e. oversamp = =
4. Values between 1 and10 were tried for the upsamp and oversamp parameters to testthe robustness of the line profiles. We find that the line profilesdo not change significantly for this range in values.
3. Results
The [C ii ] 157 µ m and [O i ] 63 µ m lines are detected at a signalto noise greater than 30 for both A1068 and A2597. The muchweaker [N ii ] 122 µ m and [O i b] 145 µ m lines are detected at the3–5 σ level for A2597. The [O iii ] 88 µ m line was not detected inA2597, an upper limit for this line is given in Table 2.The line spectra are fitted by a model consisting of; (i) a lin-ear function to determine the continuum flux, and (ii) a singlegaussian function to determine the line flux. Continuum sub-tracted line spectra are shown for the central spaxel in Fig. 1for A1068 and Fig. 2 for A2597. The fitted line centers agreewell with the redshift of CO in the BCG and the fitted FWHMline widths indicate gas with velocities of 300–500 km s − .The [C ii ] and [N ii ] lines in the central spaxel of both ob-jects are well described by a single gaussian. However, the [O i ]63 µ m lines, where the PACS spectral resolution is best, have pro-files indicative of weak (2–3 σ ) deviations from a single gaussianfunction. The [O i ] line in A1068 hints at a two-component struc-ture in the form of a narrow core component on top of a broad un-derlying component comparable to the CO(2-1) profile in Edge(2001). Both [O i ] lines observed in A2597 appear to have theirdominant flux component at the systemic redshift of the BCGand a weaker component o ff set by about +
250 km s − which isalso seen in the CO data (Salom´e, priv. comm). We attribute theshared structure of these atomic and moledular lines to gas kine-matics rather than self-absorption as the observed emission isfrom a large number of clouds that have much narrower intrinsicline width.The resolution of PACS at the observed wavelengths variesfrom about 5 ′′ for the [O i ] 63 µ m line to about 14 ′′ for the [C ii ]157 µ m line. We have investigated line emission in all 25 spaxelsof the PACS FoV. In all cases the line flux is dominated by thecentral spaxel. Summing up the flux in all 25 spectra and com-paring it to the flux in the central spaxel shows no evidence ofexcess line flux as compared to what is expected from a pointsource. In order to properly recover the full beam line fluxes wehave applied point source corrections (appendix A of the PSPCdocument) to the central spaxel integrated line fluxes. The resultsare listed in Table 2. This spatial resolution matches the best sub-mm interferometry results for CO (Edge & Frayer 2003; Salom´e& Combes 2004) which implies that most of the emission is onscales < ′′ so we believe our PACS line fluxes can be comparedto literature values without large beam corrections.
2. C. Edge et al.:
Herschel observations of FIR emission lines in brightest cluster galaxies
Table 1.
Log of
Herschel
PACS observations
Cluster Redshift Line Wavelength Obsid Bandwidth Resolution Beam Size( µ m) (km s − ) (km s − )A1068 0.1386 [C ii ] 179.61 1342186308 1200 201 13.5 ′′ /
33 kpc[O i ] 71.94 1342186307 600 55 5.4 ′′ /
13 kpcA2597 0.0821 [C ii ] 170.78 1342187125 1100 218 12.8 ′′ /
20 kpc[O i ] 68.41 1342187124 550 68 5.1 ′′ / ii ] 131.94 1342188942 1200 281 9.9 ′′ /
15 kpc[O i b] 157.56 1342188704 1200 241 11.8 ′′ /
18 kpc[O iii ] 95.61 1342188703 108 7.2 ′′ /
11 kpc
Table 2.
Spectral line results for
Herschel
PACS observations
Cluster Redshift Line Integrated Line Flux Velocity o ff set measured FWHM instrinsic FWHM(10 − W m − ) (km s − ) (km s − ) (km s − )A1068 0.1386 [C ii ] 104.7 ± + ±
55 378 ±
40 320 ± i ] 64.8 ± + ±
50 356 ±
40 352 ± ii ] 58.5 ± + ±
60 463 ±
40 408 ± i ] 54.7 ± + ±
55 411 ±
40 405 ± ii ] 3.8 ± + ±
60 578 ±
90 505 ± i b] 3.3 ± ±
65 484 ±
90 420 ± iii ] < . Fig. 1.
Herschel
PACS spectra of [C ii ] and [O i ] in A1068
4. Discussion
The primary result from the SDP observations for this project isthat the atomic cooling lines are present in both observed clus-ters. This first detection of these lines in cluster cores reinforcesthe importance of the cold gas in these environments. However,there are a number of questions that these detections raise.
How do the properties of the FIR lines compare to lo-cal LIRGs / ULIRGs?
There have been several studies of localgalaxies with ISO and high redshift galaxies using ground-based instrument that cover [C ii ] and [O i ] (Malhotra et al. 1997,Maiolino et al. 2005, Hailey-Dunsheath et al. 2010). These stud-ies show that the ratio of [C ii ] to FIR luminosity is a function ofluminosity with relatively less [C ii ] emission for the most FIR lu-minous sources. Using the FIR data from Edge et al. (2010), wecalculate the [C ii ] to FIR luminosity ratios are 10 − . and 10 − . for A1068 and A2597, respectively. The [C ii ] / FIR ratios of thesetwo galaxies are comparable to those measured for galaxies ofsimilar L
FIR (see Fig. 2 of Maiolino et al. (2005)). In particu-lar, the [C ii ] to FIR luminosity ratio is lower for the more FIRluminous of the two galaxies. The ratio of [C ii ] to [O i ] showsless variation (1.62 and 1.07) and is again consistent with othercomparable galaxies (Luhman et al. 2003). Also our CO(1-0) to FIR luminosity ratios of 10 − . and 10 − . for A1068 and A2597are consistent with star-forming local galaxies (Malhotra et al.1997). So, despite potential di ff erences in excitation, pressureand metallicity, the relative intensity of the atomic and molecularlines to the FIR luminosity do not distinguish the BCGs studiedhere from other FIR bright galaxies. How do the dynamics of atomic and molecular lines com-pare?
The relative velocity width of the atomic lines comparedto the CO and MIR H lines can provide important diagnos-tics for the dynamics and energetics of the various gas trac-ers. From the line width alone the resolution corrected lineFWHM widths for the [C ii ] and [O i ] lines are ≈
330 km s − and ≈
400 km s − for A1068 and A2597 respectively. This comparesto 243 ±
13 km s − (Edge 2001) and 292 ±
45 km s − (Salom´e,priv. comm.) for CO(2-1) for A1068 and A2597. In each casethe FIR lines are a factor of ∼ ff erent resolution. Instead, this di ff erence is more likelyto be related to the lines being emitted from di ff erent regionswithin the BCG or in shocks. However, this clearly needs to
3. C. Edge et al.:
Herschel observations of FIR emission lines in brightest cluster galaxies
Fig. 2.
Herschel
PACS spectra of [C ii ], [O i ] (63 µ m), [N ii ] and [O i b] (145 µ m) in A2597.be tested in more systems and through direct comparison of the[C ii ] and [O i ] extent with that of CO. How do the FIR line ratios constrain the gas properties?
Therelative strength of the FIR lines can constrain several key prop-erties of the gas phase that dominates the emission. The mainconstraint we can determine directly from our current data isfrom the [C ii ] 158 µ m and [O i ] 63 and 145 µ m lines for A2597.Kaufman et al. (1999) present photodissociation region (PDR)model predictions for the [O i ] 145 µ m / µ m and [C ii ] 158 µ mto [O i ] 63 µ m line ratios. Combining these two constraints forour observed [O i ] 145 µ m / µ m ratio of 0.06 ± i ]63 µ m to [C ii ] 158 µ m ratio of 0.94 ± . ± . cm − and an incident FUV flux of G of 150–1000Habing units. These values of G imply intrinsic FUV luminosi-ties of ≈ − × erg s − if the clouds subtend 3–5 kpc. This iscomparable to the observed FUV luminosities of these galaxiesonce dust absorption is taken into account (O’Dea et al. 2004).
5. Conclusions
These initial results from
Herschel indicate that atomic coolinglines are present in the brightest cluster galaxies in cooling flowclusters. The intensity and velocity width of these lines is con-sistent with all the other observed tracers of cold gas in thesesystems implying they originate from the same population ofclouds. The only apparent exception to this in our current obser-vations is that the FIR lines appear to be systematically broaderthan the CO lines impling that the relative intensity of theselines varies with position within the BCG. The results that willcome from our Open Time Key Project for 11 BCGs will expandgreatly on those presented here with more lines and a greater range of BCG properties. Beyond this, the potential for
Herschel to illuminate the properties of the cold gas that may fuel cold nu-clear accretion in more distant clusters and local groups is vast.
Acknowledgements.
We would like to thank the
Herschel
Observatory and in-strument teams for the extraordinary dedication they have shown to deliver sucha powerful telescope. We would like to thank the HSC and NHSC consortium forhelp with data reduction pipelines. J.B.R.O. thanks HSC, the
Herschel
Helpdeskand the PACS group at MPE for useful discussions. R. M. thanks the NHSC forthe HIPE tutorials.
References
Crawford, C. S., Allen, S. W., Ebeling, H., et al. 1999, MNRAS, 306, 857Donahue, M., Jord´an, A., Baum, S. A., et al. 2007, ApJ, 670, 231Edge, A. C., 2001, MNRAS, 328, 762Edge, A. C., & Frayer, D. T. 2003, ApJ, 599, L69Edge, A. C., Oonk, J. B. R., Mittal, R., et al. 2010, A&A, submittedEgami, E., Rieke, G. H., Fadda, D., et al. 2006, ApJ, 652, L21Ferland, G. J., Fabian, A. C., Hatch, N. A., et al. 2009, MNRAS, 392, 1475Hailey-Dunsheath, S., Nikola, T., Stacey, G. J., et al. 2010, ApJL, 714, L162Ja ff e, W., & Bremer, M. N., 1997, MNRAS, 284, 1Kaufman, M. J., Wolfire, M. G., Hollenback, D. G., et al. 1999, ApJ, 527, 795Luhman, M. L., Satyapal, S., Fischer, J., et al. 2003, ApJ, 594, 758Maiolino, R., Cox, P., Caselli, P., et al. 2005, A&A, 440, L51Malhotra, S., Helou, G., Stacey, G. J., et al. 1997, ApJ, 491, L27McNamara, B. R., Wise, M. W., & Murray, S. S. 2004, ApJ, 601, 173McNamara, B. R. & Nulsen, P. E. J. 2007, ARA&A, 45, 117O’Dea, C. P., Baum, S. A., Mack, J., et al. 2004, ApJ, 612, 131O’Dea, C. P., Baum, S. A., Privon, G., et al. 2008, ApJ, 681, 1035Ott, S. 2010 in ASP Conference Series, Astronomical Data Analysis Softwareand Systems XIX, Y. Mizumoto, K. I. Morita, and M. Ohishi, eds., in pressPeterson, J. R., & Fabian, A. C. 2006, PhR, 417, 1Pilbratt, G., et al. 2010, A&A, submittedPoglitsch, A., et al. 2010, A&A, submittedSalom´e, P., & Combes, F. 2003, A&A, 412, 657Salom´e, P., & Combes, F. 2004, A&A, 415, L1
4. C. Edge et al.:
Herschel observations of FIR emission lines in brightest cluster galaxies