i ",- _.Oj/_ NASA-CR-194820 V THE SOFT X-RAY HALO OF THF,, SPIRAL GALAXY _ NGC4631 Z_ _o t_ Renfi A.M. Walterbos & Michael F. Steakley New Mexico State University, Astronomy Department, Las Cruces, NM _0 I Ch x _c_ _0 I Email ID A I l,..a [email protected] O m ,,D Q. Daniel Wang 0 University of Colorado, JILA/CASA, Boulder, CO r C ,,-4 Colin.A. Norman -1" r'_ Johns Hopkins University, Dept. of Physics. & Astron., Baltimore, MD q( r O 'I"I Robert Braun Netherlands Foundation for Research in Astronomy, Dwingeloo, NL _Z X _0 ABSTRACT We present ROSAT PSPC observations of the close to edge-on spiral galaxy NGC4631. This vigorously star forming galaxy shows extented X-ray emission perpendicular to the plane, out to about 6 to 8 kpc. The spatial extent is largest at soft X-ray energies. The total X-ray luminosity of hot gas can be 0 easily supplied by star formation in the disk, and it is likely that the halo is due to outflow of hot gas from the inner disk. Spectral analysis of the X-ray data 0 C Z shows that part of the halo emission may be quite cool, well below 106K. We 0 briefly discuss implications of these results. I ",4 0 ba 1. Introduction Ever since Spitzer (1956) predicted that halos of spiral galaxies might contain hot gas, this component of the interstellar medium (ISM) has stimulated much interest. Models of the interaction of supernovae with the ISM (Cox & Smith 1974, McKee & Ostriker 1977), and of the disk-halo interface (e.g. Bregman 1980, Norman 8z Ikeuchi 1989) have demonstrated that it is quite likely for such a hot component to exist; the surprising result is actually that it has been quite difficult to find it! (e.g. Fabbiano 1989, Fabbiano & Trinchieri 1987). NGC4631, a close to edge-on Sc spiral galaxy, may be the most likely target for finding a hot halo. It has vigorous star formation, as judged from its far-infrared luminosity (Rice e¢ al. 1988) and its extensive radio halo (e.g. Hummel & Dettmar 1990). Its central 4 kpc of the disk shows a few long Ha filaments, indicative of disk-halo interaction, and possibly a thick diffuse ionized gas layer (Rand et aI. 1992, Walterbos et aI. 1993, 1994). In addition, the galaxy is only 7.5 Mpc distant, and its high Galactic latitude implies a low foreground HI column. The latter is crucial for the detection of X-rays below 0.4 keV, the range where hot gas is expected to emit most of its energy. However, NGC4631 may not be a proto-typical late-type spiral because it is interacting with two neighboring galaxies, producing extensive HI tidal tails (Rand & van der Hulst 1993), a radio halo, and enhanced star formation in the disk. 0 20 40 32 37- 36-- 35-- O o 34-- O z 33-- 2 t.- '< 32-- z m _1 ouJ 31-- © Q 29 28 I t 12 42 45 30 15 00 41 45 30 RIGHT ASCENSION (J2000) 0 20 40 - 30 15 00 41 45 30 RIGHT ASCENSION (J2000) Figure I. (Top) PSPC image of NGC4631 in the range 0.2-0.4 keV, super- posed on an Halpha image (grey scale; note that the spatial alignment is still somewhat uncertain). Contours at i, 2, 3, 4, and 5 (times 10(-6) counts/sec/pix) above background. (Bottom) The X-ray emission in the range 0.5-0.9 keV. Contours start at 2 and increase in steps of 2 (same units). One arcmin corresponds to 2.2 kpc. 2. Observations and results We obtained a 23 ksec ROSAT PSPC observation of NGC4631, of which about 16 ksec was left after editing the data to delete periods with enhanced non-cosmic background count rates (Snowden et al. 1994). Complete results of this observation are presented by Wang et al. (1993) and by Walterbos et al. (1994). Data below 0.2 keV suffer badly from the electronic ghost problem, so these could not be used in creating images, but only in the spectral analysis. About 1200 counts were detected, 40% each from disk and halo, and 20% from a bright source west of the nucleus (Pig. 1) also seen by Einstein (Fabbiano £c Trinchieri 1987). This source appears to coincide with a giant HI supershell (Wang et al. 1993). We plot X-ray contours over an Ha image in Fig. 1 for the soft and medium energy ranges. The X-ray emission, although weak (the highest contour in Fig. 1 (top) is about 5 sigma), is clearly extended above the disk, out to about 6 to 8 kpc. The extra-planar X-ray emission is prominent above the region of most intense Ha emission, near the "double worm" in Ha (Rand et al. 1992). The halo emission appears asymmetric, which may be intrinsic or due to the orientation at which we see NGC4631 (Walterbos et al. 1994). The medium energy image is also significantly extended in z-direction. The hard X-ray image (not shown) is more confined to the disk, which implies that the extra-planar medium X-rays are also likely due to hot gas, not to a population of hard X-ray binaries. 3. Spectral analysis and discussion The spectra for the halo and disk regions are dramatically different (Fig. 2). A two-component fit to the disk spectrum is required, with one component hotter than 2.4 x 107K, probably due to hard X-ray binaries, and a cooler com- ponent of 3.5 x 106K, which we tentatively identify with hot gas in the disk. A foreground column of 1021 atoms cm -2 was assumed in the fit; the derived temperatures are not extremely sensitive to this assumption. The halo spectrum appears more complicated. A single temperature fit (middle panel) works well, and indicates a temperature of 2.8 x 106K, similar to the temperature of hot gas in the disk. The total luminosity for the hot gas (disk and halo) is about 5 - 7 x 10a9 erg s-1. This luminosity is easily supplied by the energy input from supernovae, which amounts to about 1042 erg s-1. The hard disk component contributes about 4 x 10a9 erg s-1. Although the single temperature fit to the halo gas is appealing there may be a problem: the fit requires a very low foreground column of less than 3.3 x 1019 atoms cm -2, which does not agree with the observed foreground HI column of 1.2 x 102o atoms cm -2. There is likely also a substantial Galactic H+ foreground column (e.g. Reynolds 1991). If we assume a foreground column of 1.8 x 1020 atoms cln -2, a single temperature fit to the halo spectrum is rejected at the 95% confidence level (top panel). In that case a significant fraction of the halo gas must be cool, about 3 x 10SK. Because emission from this gas would be so heavily absorbed, the excess in Fig. 2 would actually imply an order of magnitude larger (in comparison to the several million K gas) X-ray luminosity. If correct, this might be where a large fraction of the supernovae energy needs to go. It would also explain the previous lack of success in detecting hot gas in spirals. Only the very favorable conditions for NGC4631 give us a hint of this > o C o y o 0 o 0.1 0.2 0.5 1 2 > ° ] _ N - o C d 0 o .1 0,2 0.5 1 2 o d 0 . - ! 0 0 i 0.1 0.2 0.5 1 2 Chonnel energy (keV) Figure 2. S_ctral fits to t_ halo (top and md_e)_d diskX-ray data (bottom). The two halo fits corres_ to the two options diseased in the t_t; t_ halo emission is clearly quite soft. _e disk s_ctrum is _rder, _rtially _e to absorption but also to a contri_tion of hard X-ray so_ces. intriguing option andunfortunately it may behard to find othergalaxieswhere ROSAT data could confirm the presenceof large amountsof "cool" hot gas. A gratifying result is that wehavedetectedX-ray emissionfrom hot gas in both the disk and halo of NGC4631,confirming a long-suspectedISM com- ponent, andmodelsof disk-halointeraction (Bregman 1980,Norman& Ikeuchi 1989).Onthe othd hand, the strong tidal interaction doesnot makeNGC4631 a typical spiral, so extrapolation of our results to other galaxiesmay not be entirely appropriate. Outflows of hot gashavebeenseenin starburst galaxies (seeBregmanandHeckman,theseproceedings)andalthoughNGC4631hasnot typically beenplacedin this class,the concentrationofhalogasnearthe central disk might indicate that alsohereit is especiallythe central(but not necessarily nuclear)regionthat is responsiblefor poweringthehalo. Thusit remainsuncer- tain howmany "normal" spiralswill havehot halos. Theother likely candidate, NGC891 (Bregman,theseproceedings,and Bregman&:Pildis 1993)doesshow a thick disk of hot gas,but not asextensiveas NGC4631,possiblybecauseof the largerforegroundcolumn which makesit impossibleto seeavery soft X-ray halo. Apart from thesetwo systems,whichboth showhighstarformation rates, evenROSATmayhaveahard time detectingextendedsoftX-ray halosin other "normal" (i.e. non-starburst)edge-on spirals,although hot gasin the diskswill likely be found. Acknowledgements We are grateful to Richard Rand for making his integrated HI imageof NGC4631availableto us. This work hasbeensupportedby grantsfrom NASA (NAG 5-1924)andNSF (AST9123777)to RAMW. References Bregman,J.N., 1980,ApJ 236,577 Bregman,J.N., Pildis, R.A., 1993,ApJ , in press Cox, D.N., Smith, B.W., 1974,ApJL 189,L105 Fabbiano,G., 1989,ARA&A 27, 87 Fabbiano,G., Trinchieri, G., 1987,ApJ 315,46 Hummel,E., Dettmar, R.-J., 1990,A&A 236,33 McKee, C.F., Ostriker, J.P., 1977,ApJ 218,148 Norman, C.A., Ikeuchi, S., 1989,ApJ 345,372 Rand,R.J., Kulkarni, S.R.,Hester,J.J.: 1992,ApJ 396,97 Rand,R.J., vander Hulst, J.M., 1993,AJ 105,2098 Reynolds, R.J., 1991,in The Interstellar Disk-Halo Connection,IAU Syrup. 144,ed. J.B.G.M. Bloemen,(Dordrecht: Kluwer), p 67 Rice,W.A., Lonsdale,C.J., Soifer,B.T., Neugebauer,G., Koplam E.L., Lloyd, L.A., de Jong,T., Habing,H.J., 1988,ApJS 68,91 Snowden,S.L., McCammon,D., Burrows,D.N., Mendenhall,J.A., 1994,ApJ ,in press Spitzer, L., 1956,ApJ 124,20 Walterbos, R.A.M., Braun, R., Norman, C.: 1993,in Evolution ofGalaxiesandTheir Environment, eds.Hollenbach,Thronson,& Shull,NASA Conf. Publ. 3190,p326 Walterbos,R.A.M., Steakley,M.F., Wang,Q.D., Norman, C.A., Braun,R., 1994,to be submitted to ApJ Wang, Q.D., Walterbos, R.A.M., Steakley,M.F., Norman, C.A., Braun, R., 1993,sub- mitted to Nature