1: % mnsample.tex
2: %
3: % v1.2 released 5th September 1994 (M. Reed)
4: % v1.1 released 18th July 1994
5: % v1.0 released 28th January 1994
6:
7: \documentclass{mn2e}
8: \voffset-.8in
9: %\usepackage{natbib}
10: %\bibliographystyle{abbrvnat}
11: %\citestyle{mn}
12: % If your system has the AMS fonts version 2.0 installed, MN.sty can be
13: % made to use them by uncommenting the line: %\AMStwofontstrue
14: %
15: % By doing this, you will be able to obtain upright Greek characters.
16: % e.g. \umu, \upi etc. See the section on "Upright Greek characters" in
17: % this guide for further information.
18: %
19: % If you are using AMS 2.0 fonts, bold math letters/symbols are available
20: % at a larger range of sizes for NFSS release 1 and 2 (using \boldmath or
21: % preferably \bmath).
22:
23: \newif\ifAMStwofonts
24: \AMStwofontstrue
25: \usepackage{graphicx}
26: %%%%% AUTHORS - PLACE YOUR OWN MACROS HERE %%%%%
27: \newcommand{\etal}{et~al.~}
28:
29:
30: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
31:
32:
33:
34:
35: \title{X-ray Emission from Haloes of Simulated Disc Galaxies}
36: \author[S. Toft et al.]
37: {S. Toft,$^1$\thanks{E-mail: toft@astro.ku.dk},
38: J. Rasmussen$^1$, J. Sommer-Larsen$^2$ and K. Pedersen,$^1$\\
39: $^1$Astronomical Observatory, Copenhagen University, Juliane Maries Vej 30, DK-2100 Copenhagen \O, Denmark \\ $^2$Theoretical Astrophysics Center, Juliane Mariesvej 30, DK-2100 Copenhagen \O, Denmark }
40: \date{}
41:
42: \pagerange{\pageref{firstpage}--\pageref{lastpage}}
43: \pubyear{2001}
44:
45: \begin{document}
46:
47: \maketitle
48:
49: \label{firstpage}
50:
51: \begin{abstract}
52: Bolometric and 0.2-2 keV X-ray luminosities of the hot gas haloes of simulated
53: disc galaxies have been calculated at redshift $z$=0. The TreeSPH simulations are
54: fully cosmological and the sample of 44 disc galaxies span a range in
55: characteristic circular speeds of $V_c$ = 130-325 km s$^{-1}$. The galaxies have been
56: obtained in simulations
57: with a considerable range of physical parameters, varying the baryonic
58: fraction, the gas metallicity, the meta-galactic UV field, the cosmology, the
59: dark matter type, and also the numerical resolution. The models are found to
60: be in agreement with the (few) relevant X-ray observations available
61: at present. The amount of hot gas in the haloes is also consistent
62: with constraints from pulsar dispersion measures in the Milky Way.
63: Forthcoming XMM and Chandra observations should enable much more stringent
64: tests and provide constraints on the physical parameters.
65: We find that simple cooling flow models over-predict X-ray luminosities
66: by up to two orders of magnitude for high (but still realistic)
67: cooling efficiencies relative to the models presented here.
68: Our results display a clear trend that {\it increasing} cooling efficiency
69: leads to {\it decreasing} X-ray luminosities at $z$=0. The reason is found
70: to be that increased cooling efficiency leads to a decreased fraction of
71: hot gas relative to total baryonic mass inside of the virial radius at
72: present. At gas metal
73: abundances of a third solar this hot gas fraction becomes as low as
74: just a few
75: percent.
76: We also find that most of the X-ray emission comes from the inner
77: parts ($r\la 20$ kpc) of the hot galactic haloes.
78: Finally, we find for realistic choices of the physical parameters
79: that disc galaxy haloes possibly were {\it more than one order of magnitude}
80: brighter in soft X-ray emission at $z$$\sim$1, than at present.
81:
82:
83:
84: \end{abstract}
85:
86: \begin{keywords}
87: methods: N-body simulations --cooling flows --galaxies: evolution --galaxies: formation --galaxies: halos --galaxies: spiral --X-rays: galaxies
88:
89: \end{keywords}
90:
91: \section{Introduction}
92: In disc galaxy formation models infall of halo gas onto the disc due
93: to cooling is a generic feature. However, the gas accretion rate and
94: hot gas cooling
95: history are at best uncertain in all models so far. It is thus not clear to
96: which extent
97: the gas cooling out from the galaxy's halo is replenishing that which is
98: consumed by star formation in the disc.
99: Such continuous gas infall is essential to
100: explain the extended star formation histories of isolated spiral
101: galaxies like the Milky-Way and the most likely explanation of the ``G-dwarf
102: problem'' --- see, e.g., Rocha-Pinto \& Marciel (1996) and Pagel (1997).
103:
104: At the virial temperatures of disc galaxy haloes the dominant cooling
105: mechanism is thermal bremsstrahlung plus atomic line emission.
106: The emissivity, increasing strongly with halo gas density, is expected to
107: peak fairly close to the disc and decrease outwards,
108: and if the cooling rate is significant
109: the X-ray flux may be visible well beyond the optical radius of a galaxy.
110:
111: Recently, Benson \etal (2000) compared ROSAT observations
112: of three
113: massive, nearby and highly inclined disc galaxies with predictions of simple
114: cooling flow models of galaxy formation and evolution.
115: They showed that these models predict about {\emph{an order of
116: magnitude}} more X-ray emission from the galaxy haloes (specifically
117: from a 5-18 arcmin annulus around the galaxies) than observational detections and upper limits.
118:
119: In this paper we present global X-ray properties of the haloes of a large,
120: novel sample of model disc galaxies at redshift $z$=0. The galaxies result from
121: physically realistic gravity/hydro simulations of disc galaxy
122: formation and evolution in a cosmological context. We find that our model
123: predictions of X-ray properties
124: of disc galaxy haloes are consistent with observational detections and
125: upper limits. Given the
126: results of the theoretical models of
127: Benson \etal we list the most important reasons why simple cooling flow models
128: over-predict the present day X-ray emission of disc galaxy haloes.
129:
130: In section 2 we give a very short description of the disc galaxy simulations.
131: In section 3 we briefly describe the X-ray halo emission calculations and in section 4 the
132: results obtained. Section 5 constitutes the discussion and section 6
133: the conclusion.
134: \begin{figure}
135: \resizebox{\hsize}{!}{{\includegraphics{Tcut.ps}}}
136: \caption{The figure shows the temperature of the SPH gas
137: particles in a typical disc galaxy from our simulations versus their x
138: coordinate (one of the axes in the disc). The ``cold'' (log$T<4.5$) gas which
139: is primarily situated
140: in the disc is removed from the catalogues since it does not
141: contribute to the X-ray flux. }
142: \label{figT}
143: \end{figure}
144:
145: \section{Disc galaxy simulations}
146: We have in recent years developed novel models of formation and evolution of
147: disc galaxies. The model disc galaxies result from {\it{ab initio}},
148: fully cosmological ($\Omega_M=1$ or $\Omega_M+\Omega_{\Lambda}=1$),
149: gravity/hydro simulations. These simulations are started at a
150: sufficiently high redshift ($z_i$=20-40) that the density
151: perturbations are still linear and are then evolved through the entire
152: non-linear galaxy formation regime to the present epoch ($z$=0). The
153: code uses a gridless, fully Lagrangian, 3-D TreeSPH code incorporating
154: the effects of radiative cooling and heating (including the effects of
155: a meta-galactic UV field), inverse Compton cooling, star formation, and
156: energetic stellar feedback processes.
157:
158: A major obstacle in forming realistically sized disc galaxies
159: in such simulations is the so-called ``angular momentum
160: problem'' (e.g., Navarro \& White 1994, Sommer-Larsen et al. 1999).
161: We overcome this problem in two different ways: a) Using
162: cold dark matter (CDM) + stellar feedback processes
163: \cite{sommerlarsen02} or b) Using warm dark matter (WDM)
164: \cite{sommerlarsendolgov}. A total of 44 such disc galaxy
165: models with characteristic circular velocities in the range
166: $V_c$ = 130--325 ${\rm km\,s^{-1}}$ form the basis of the predictions
167: presented in this paper.
168: The simulations initially consist of 30000-400000 SPH+DM particles and
169: in the majority of them some of the SPH particles are turned into star
170: particles over the course of the simulation --- for details
171: about the TreeSPH simulations we refer the reader to the above quoted
172: references.
173:
174: Most of the simulations were run with primordial gas composition (76\% H and
175: 24\% He by mass) under the
176: assumption that the inflowing, hot gas is fairly unenriched in heavy elements.
177: To test the effects of metal abundance eight $\Lambda$CDM simulations
178: (four with ``universal'' baryonic fraction $f_b$=0.05 and four with $f_b$=0.10)
179: were run with a gas abundance of 1/3 solar (specifically [Fe/H]=-0.5). This is
180: the metal abundance of the intracluster medium and can probably be considered
181: a reasonable upper limit to the metal abundance of the hot gas in disc galaxy
182: haloes.
183:
184: \begin{figure}
185: \resizebox{\hsize}{!}{{\includegraphics{LxVc.ps}}}
186: \caption{Bolometric luminosity as a function of characteristic circular
187: speed.
188: {\bf Small symbols}: Flat $\Omega_M=1.0$ cosmology:
189: Open symbols: baryon fraction $f_b$=0.05, filled circles
190: $f_b$=0.1. Triangles: without UV field, non-triangles: with a UV field
191: of the Efstathiou (1992) type.
192: Connected symbols are the same galaxies run with medium (open
193: circles) and high (open circles with crosses) resolution. All
194: simulations represented by small symbols have primordial abundance.
195: {\bf Large symbols}: Flat ($\Omega_{\Lambda},\Omega_M)=(0.7,0.3)$ cosmology:
196: Open symbols: $f_b$=0.05, filled symbols $f_b$=0.1. Circles correspond to
197: primordial abundance and with a Haardt \& Madau (1996) UV field, squares
198: correspond to $Z=1/3~Z_{\odot}$ (using the cooling function of Sutherland \&
199: Dopita 1993, which does not include effects of a UV field).
200: The curves are the $L_{X,bol}$-$V_c$ relationship for the simple cooling flow
201: models (for $\Lambda$CDM NFW haloes) described in
202: Sec. \ref{models}.
203: The curves represent different baryonic fractions
204: (solid curves have $f_b=0.1$, dotted curves have $f_b=0.05$) and
205: abundances (thick curves: primordial abundances, thin
206: curves: $Z=1/3~Z_{\odot}$). }
207: \label{lxvc}
208: \end{figure}
209:
210:
211:
212: \begin{figure}
213: \resizebox{\hsize}{!}{{\includegraphics{Lx0.2-2Vc.ps}}}
214: \caption{0.2-2 keV band luminosity as a function of characteristic circular
215: speed. Symbols as in Fig. 2.}
216: \label{lx0.2-2vc}
217: \end{figure}
218:
219:
220:
221:
222:
223: \section{X-ray halo emission calculation}
224: For each of the simulated galaxies at $z$=0 we create a catalogue of SPH gas
225: particle positions, densities, temperatures and masses.
226: For each catalogue a box of size (1000 kpc)$^3$ centered on
227: the galaxy is
228: retained and all gas particles with temperatures log(T)$<4.5$ are cut away
229: (see Fig. \ref{figT}) since they will not contribute to
230: the X-ray flux (these are mainly gas particles which have cooled onto the
231: disc). The density and temperature of each particle is then averaged over
232: itself and its five nearest neighbors using a spherical smoothing kernel
233: proposed by Monaghan \& Lattanzio (1985)\footnote{It is important to note that we average
234: over the {\it original} densities from the TreeSPH simulations.
235: One can show that if the cold gas particles are cut away and the
236: densities of the remaining gas particles are then recalculated using
237: the full SPH procedure the
238: X-ray luminosities will be underestimated due to resolution problems at the
239: disc--halo interface. },
240: and a volume is assigned to the particle
241: given its mass and density. Using the average
242: temperatures and densities, each SPH particle is treated as an optically
243: thin thermal plasma, and the associated X-ray luminosity is
244: calculated at the relevant position in a given photon energy
245: band with the {\em meka} plasma emissivity code \cite{mewe}.
246: X-ray luminosities are then computed by summation over all particles in the
247: volume of interest. The bolometric X-ray luminosity is calculated using the
248: $0.012-12.4$ keV band.
249:
250: \begin{figure}
251: \resizebox{\hsize}{!}{{\includegraphics{surfbrightpaper.ps}}}
252: \caption{Mean bolometric X-ray surface brightness
253: profiles perpendicular to the plane of the disc of three high $V_c$
254: galaxies (solid curve) and four Milky Way sized galaxies (dashed
255: curve). The physically most important parameters for these galaxies are $f_b$=0.10 and
256: primordial gas abundance.
257: The profiles were calculated by binning the emission in
258: 40 kpc wide, 5 kpc high slices parallel to the disc.}
259: \label{surfbright}
260: \end{figure}
261:
262: \section{Results}
263: In Fig. \ref{lxvc} the total bolometric X-ray luminosities $L_{X,bol}$ of the 44 simulated disc galaxies in our sample are plotted versus their
264: characteristic circular speed $V_c$, defined as the circular velocity in the
265: disc at $R_{2.2}= 2.2~R_d$, where $R_d$ is the disc scale length -- see
266: Sommer-Larsen \& Dolgov (2001) for details.
267: The X-ray luminosities derived from the
268: simulations are up to two orders of magnitude
269: below values derived from simple cooling flow models which are
270: described in Sec. 5.1. As can be seen from the
271: figure $L_{X,bol}$ $\sim$
272: $10^{40}$ erg s$^{-1}$ for a Milky Way sized galaxy.
273:
274: As expected from the simple models, the simulated galaxies display an
275: $L_{X,bol}$-$V_c$ relation, but with a significant scatter. Part of this
276: scatter arises from the different
277: conditions under which the simulations have been run (see
278: Sec. \ref{robustness}) and part of it is a ``real'' scatter arising
279: from the different geometries and cooling histories of the individual galaxy
280: haloes.
281: This ``real'' scatter can be estimated by inspection of the large filled circles
282: in Fig. 2 which represent simulations run with the same, physically important
283: parameters (see the figure caption). These data points display a rms
284: dispersion of about 50\% around the mean.
285:
286: There is a tendency for galaxies formed in simulations with baryon fraction
287: $f_b$=0.05 and primordial gas abundance to have systematically higher
288: $L_{X,bol}$ (by about a factor of two) than galaxies formed in similar
289: simulations with $f_b$=0.10. Also, galaxies formed in simulations with gas
290: abundance $Z=1/3~Z_{\odot}$ tend to have systematically lower $L_{X,bol}$
291: than the ones with primordial gas. We discuss these trends in Sec.
292: \ref{models}.
293:
294: Fig. 3 shows the 0.2-2 keV band X-ray luminosities of the 44 sample
295: disc galaxies versus $V_c$. The systematic trends mentioned above are also
296: seen in this plot, in particular is the difference between the $f_b$=0.05 and
297: 0.10 simulations (with primordial gas) even more pronounced than in
298: Fig. 2. We also discuss this in Sec. \ref{models}.
299:
300: Most (but not all) of the X-ray emission originates from regions of
301: the hot gas halo fairly close to the disc: 95\% of the emission typically originates
302: within about 20 kpc of the
303: disc. This is illustrated in Fig.~\ref{surfbright} where we plot the
304: mean surface brightness profiles perpendicular to the disc of
305: three high $V_c$ galaxies and four Milky Way sized galaxies.
306:
307: \section{Discussion}
308: \label{models}
309: \subsection{Comparison to simple cooling flow models}
310: In order to compare our results with previous work, we
311: calculated a family of simple cooling flow models similar to those
312: considered by Benson \etal (2000).
313:
314: In these models it is assumed that the cooling occurs in a static potential
315: and that the gas initially was in place and traced the dark matter (DM).
316: Gas is assumed to flow from the cooling radius (at which the cooling
317: time equals the age of the universe) to the disc (settling there as cold gas)
318: on a time-scale much shorter than the Hubble time. The bolometric X-ray
319: luminosity can
320: then be approximated simply as the mass accretion rate,
321: $\dot{M}_{ \rm{cool}} = 4 \pi r_{\rm{cool}}^2 \rho_{\rm{gas}}(r_{\rm{cool}})
322: ~\dot{r}_{\rm{cool}}$, times the gravitational potential difference, so
323: \begin{equation}
324: L_{{\rm X,bol}}=\dot{M}_{{\rm{cool}}}(r_{{\rm{cool}}})\int^{{r_{\rm cool}}}_{{r_{\rm
325: optical}}}
326: \frac{V_c^2(r)}{r}{\rm d}r.
327: \end{equation}
328:
329: In Fig. \ref{lxvc} we show for different baryonic fractions and halo gas
330: metallicities the $L_{X,bol}$-$V_c$ relationships expected from such
331: simple models for haloes of the $\Lambda$CDM NFW types \cite{NFW} - the
332: relationships for SCDM NFW haloes are very similar. Effects of a meta-galactic
333: UV field were not included in these simple models --- such effects are less
334: significant than the effects of varying the baryon fraction and the
335: gas abundance.
336:
337: The disc galaxies resulting from our simulations follow an
338: $L_{X,bol}$-$V_c$ relation
339: with approximately the same slope as the relation derived from
340: simple cooling flow models ($L_{X,bol} \propto V_c^5$ - see Fig. 2), but
341: shifted to systematically lower $L_{X,bol}$ and with a significant scatter.
342: At small $V_c$ the relation looks steeper, but
343: this can be attributed to the systematically lower $L_{X,bol}$ of galaxies
344: formed in simulations with a uniform gas abundance
345: %(i.e. also in all the hothalo gas)
346: of 1/3 solar rather than primordial -- see sec. \ref{metal}.
347:
348:
349: %At velocities $v_c<220$ kms$^{-1}$ the relation might be a bit
350: %steeper.
351: %This is most likely due to the fact that the cooling radius exceeds
352: %the virial radius of these galaxies. For the semi-analytical models, the halo was extended beyond $r_{virial}$ allowing
353: %gas outside the virial
354: %radii to cool. This is unphysical and does not happen in
355: %our more realistic simulations. The steepening of the $L_X$-$v_c$
356: %at $v_c<220$km s$^{-1}$ could thus be connected to the exhaustion of the
357: %virial volume hot gas reservoir.
358:
359: %At velocities $v_c>220$kms$^{-1}$ the slope of the $L_X$-$v_c$
360: %relation of the simulated galaxies approaches the slope predicted
361: %by semi-analytical models ($L_X \propto v_c^5$, ref!). At lower
362: %velocities it is steeper.
363: %This is most likely due to the fact that the cooling radius exceeds
364: %the virial radius of these galaxies. For the semi-analytical models, the halo was extended beyond $r_{virial}$ allowing
365: %gas outside the virial
366: %radii to cool. This is unphysical and does not happen in
367: %our more realistic simulations. The steepening of the $L_X$-$v_c$
368: %at $v_c<220$km/s could thus be connected to the exhaustion of the
369: %virial volume hot gas reservoir.
370:
371:
372: The main reasons why the simple cooling flow models predict up to two orders
373: of magnitude more X-ray luminosity than the physically more realistic
374: simulated galaxies considered here are:
375:
376: \begin{enumerate}
377: \item The assumption that the DM potential is static and that all gas is in
378: place at the beginning of the
379: galaxy's history, tracing the assumed static DM halo. In our
380: simulations the distribution of gas in the DM haloes is dynamic in the sense
381: that it changes due to discrete merging and interaction events,
382: especially in the early phases of the simulations, $z\ga$1-2.
383: \item The assumption that gas particles at the cooling radius instantly
384: cool out from a static infinite reservoir of hot gas. The gas halo
385: profile is continuously changing everywhere in the halo as the hot gas
386: cools out, not just at the cooling radius. Moreover, only the hot gas inside
387: of the virial radius is available for disc formation --- the effect of
388: this in cases of high cooling efficiency is very significant as
389: discussed in section 5.3.
390: \item The bulk of the cooling and X-ray emission is determined by what is
391: going on fairly close to the disc ($r \la 20$kpc) rather than what is
392: happening at the cooling radius (see Fig.\ref{surfbright}).
393: \end{enumerate}
394:
395: %One can show that the rate at which hot halo gas cools out is proportional
396: %to $\frac{L_{{\rm X,bol}}}{<T>}$, where $\frac{1}{<T>}$ is the emissivity
397: %weighted inverse temperature of the hot gas. Hence the physical reason for
398: %the significantly lower X-ray luminosities of the present models compared to
399: %the simple cooling flow ones is that the points listed above in combination
400: %lead to
401: %present day gas accretion rates which are substantially lower for our models
402: %($\sim$0.5 M$_{\odot}$/yr for Milky Way sized disc galaxies -- see Sommer-Larsen
403: %\etal 2002) than for the simple cooling flow models.
404: %(5-6 M$_{\odot}$/yr -- see Benson \etal 2000)
405:
406:
407: \begin{figure}
408: \resizebox{\hsize}{!}{{\includegraphics{lxappplot_dist13.8000_paper1.ps}}}
409: \caption{Bolometric luminosity in the annulus considered by Benson \etal
410: (2000) (a ring around the galaxy, with an
411: inner radius of 5 arcmin and an outer radius of 18 arcmin) of all
412: the disc galaxies in our simulation sample (assuming a distance of 13.8 Mpc).
413: All galaxies are viewed ``edge on'', the error bars represents the
414: difference in $L_X$ resulting from viewing the galaxies along the x and y
415: directions, respectively. The observed data point
416: (indicated by an asterisk) and its error bars are adopted from Benson \etal
417: Other symbols as in Fig. 2.}
418: \label{lxappdist}
419: \end{figure}
420:
421: \subsection{Comparison to observations}
422: When trying to determine the amount of X-ray halo emission observationally,
423: it is important not to
424: confuse the signature of cooling, hot gas with the emission from X-ray
425: binaries in the disc, the diffuse emission of the
426: galaxy's disc \cite{read}, or with the X-ray emission from hot bubbles
427: of outflowing gas that is seen in starburst galaxies (see Pietch \etal 1997
428: and Strickland \& Stevens 2000).
429:
430: In a recent deep Chandra
431: observation of a nearby edge-on spiral galaxy NGC 4631 soft X-ray
432: emission extending up to 7 kpc from the disc was observed
433: \cite{wang}. However, this disc is disturbed due to interaction and
434: the off--disc emission was attributed to processes in the disc.
435:
436: Benson \etal (2000) attempted to detect diffuse halo emission in
437: deep ROSAT observations of the haloes of three large and nearly edge-on,
438: isolated and undisturbed disc
439: galaxies; NGC 2841, NGC 4594 and NGC 5529. This study was
440: limited by the sensitivity, and poor point source removal
441: possible with ROSAT, but despite this the authors were able to
442: place limits on the X-ray emission in an annulus around the galaxies
443: (with an inner radius of 5~{arcmin} and an outer radius of
444: 18~{arcmin}). Of the three galaxies considered
445: the strongest constraint was derived for the halo of the relatively
446: nearby (distance $13.8 \pm 5.2$ Mpc) massive ($V_c=317\pm2$~km
447: s$^{-1}$) disc galaxy NGC 2841. In the annulus around this galaxy a
448: bolometric luminosity of $L_X=1.6\pm1.3 \times 10^{40}$erg s$^{-1}$
449: was derived.
450: This is much less than expected from the simple cooling-flow models.
451: Benson \etal (2000) predict (assuming $f_b$=0.06 and $Z$=0.3 $Z_{\sun}$) the luminosity in this annulus to be about 30 times higher;
452: $L_X=45\pm4.4\times 10^{40}$erg s$^{-1}$ for NFW haloes (and even more for
453: isothermal sphere haloes). However, the above result is in excellent agreement
454: with the expectation from our simulations, as illustrated in
455: Fig. \ref{lxappdist} where we plot the bolometric luminosity in the
456: considered annulus (assuming a distance of 13.8 Mpc) of all the galaxies in
457: our simulation sample versus $V_c$. We note that our three ``data'' points at
458: $V_c$=300-325 km s$^{-1}$ are for simulations with baryon fraction
459: $f_b$=0.10 and primordial gas abundance. For a {\it given} $V_c$ we would expect the bolometric luminosities to be about twice as much for
460: simulations with $f_b$=0.05 - see section 5.3.4. Such a low $f_b$ is
461: unlikely given determinations of the ``universal'' baryon fraction,
462: $f_b \approx 0.10$~$(h/0.7)^{-3/2}$, derived from galaxy clusters
463: \cite{EF} and
464: galaxy clustering \cite{per}, but in any case the
465: X-ray luminosities would still be consistent with the NGC 2841 measurement.
466: Moreover, the (more realistic) inclusion of some level of enrichment
467: in the gas and the use of the more realistic meta-galactic UV field of
468: Haardt \& Madau (1996) would tend to {\it lower} the X-ray luminosities.
469: We also note that the total X-ray luminosities of our three galaxies
470: at $V_c$=300-325 km s$^{-1}$ are ``only'' a factor of 3-5 lower than
471: the predictions of the simple cooling flow models of Benson \etal and
472: ours (the latter for $f_b$=0.10 and primordial gas). Hence an
473: important part of the reason why our models match the observed
474: bolometric luminosity of the 5-18 arcmin annulus around NGC 2841 is
475: that most of the X-rays are emitted from the inner 20 kpc of our model
476: disc galaxy haloes (5 arcmin correspond to 20 kpc at a distance of
477: 13.8 Mpc). In other words our ``geometric correction'' is considerably
478: larger than the one used by Benson \etal
479: %All galaxies are viewed ``edge on'', the error bars represents the
480: %difference in $L_X$ resulting from rotating all the galaxies 90{\degr}
481: %about their spin axis before applying the annulus.
482: %The observed data point and its error bars are adopted from \cite{benson}.
483:
484: The observational constraints on the bolometric luminosity of NGC
485: 4594 and NGC 5529 are weaker than for NGC 2841.
486: The upper limits on these are again about an order of magnitude less than
487: expected
488: from the simple models, but in agreement with the expectation
489: from our simulations.
490:
491: The diffuse X-ray luminosity of the
492: Milky Way's hot halo has recently been estimated: Pietz \etal
493: (1998) estimate a 0.1-2 keV luminosity of 7$\cdot$10$^{39}$erg s$^{-1}$ and
494: Wang (1997) a 0.5-2 keV luminosity of 3$\cdot$10$^{39}$erg s$^{-1}$. Assuming
495: a temperature of 0.15 keV (Georgantopoulus \etal 1996; Parmar \etal
496: 1999) this translates into 0.2-2 keV
497: luminosities of 5 and 7$\cdot$10$^{39}$ erg s$^{-1}$ respectively. These
498: estimates (which should probably be seen as upper limits)
499: are consistent with
500: our findings from the simulations for $V_c \simeq$ 220 km s$^{-1}$ --- see Fig. 3.
501:
502: \subsection{Effects of numerical resolution and physical parameters}
503: \label{robustness}
504: The galaxies in our sample have been compiled from a number of
505: simulations which have been run with different
506: cosmological and environmental parameters. These in-homogeneities in our
507: simulation sample may introduce
508: some scatter in the $L_X-V_c$ diagram. On the other hand,
509: this allows us to investigate trends when varying the physical parameters.
510:
511: In general, varying a parameter has impact on the present day X-ray
512: luminosity of a given galaxy if it significantly alters the ability of
513: the hot halo gas to cool during its life time.
514: In the simple cooling flow models, increasing the
515: cooling efficiency leads to an increase in $L_X$, while this is not
516: necessarily the case in more realistic simulations.
517: If the cooling efficiency is increased, more gas
518: has cooled out on the disc at $z=0$. This results in an increase of the
519: characteristic circular speed of the disc as its dynamics become more
520: baryon (and less DM) dominated, and usually a
521: {\it decrease} in the X-ray luminosity of the halo since there is less hot
522: gas left in the halo to cool and contribute to the X-ray emission.
523:
524: In the following we briefly discuss how the derived results depend on
525: the resolution of the simulation, the
526: presence of an external UV field, the assumed gas metallicity, the
527: baryon fraction $f_b$, the cosmology, the dark matter type and whether
528: or not star formation is incorporated in the simulations.
529:
530:
531: \subsubsection{Effects of resolution}
532: An important test of all numerical simulations is to check whether the
533: results depend on the resolution. This can be done from
534: Fig. \ref{lxvc} by comparing the connected symbols. These represent
535: the same galaxy, run with normal (open
536: symbols) and 8 times higher mass + 2 times higher
537: force resolution (open symbols with
538: crosses). It can be seen that this significant increase in resolution only leads to
539: a very modest increase of 19$\pm$64~\% in the X-ray luminosity relative to the
540: mean $L_X-V_c$ relation (i.e. taking the effect of the change of $V_c$ with
541: resolution into account). A similar result is inferred from Fig. 3.
542:
543:
544: \subsubsection{Effects of a meta-galactic UV field}
545: The main effects of a hard UV photon field is to ionize the gas, significantly
546: reducing its ability to cool by collisional excitation
547: (line-cooling) mechanisms \cite{vedel}.
548: This effect is evident in
549: Fig. \ref{lxvc} where the set of 4 small triangles and the set of 4
550: small open
551: circles represent the same 4 haloes, run under exactly the same conditions,
552: except that for the latter effects of a meta-galactic, redshift-dependent UV
553: field were included in the cooling/heating function. There is a
554: tendency for the galaxies without an external
555: UV-field to have lower $L_{X,bol}$ and higher $V_c$ than the galaxies
556: with external UV-field: The former have a median bolometric luminosity
557: of 55 $\pm$ 16\% of the latter (again taking into account the change
558: of $V_c$).
559:
560: So in this case an increase in the
561: cooling rate leads to a decrease in $L_{X,bol}$ at the present epoch since a smaller
562: amount of hot gas is left in the halo to produce the emission -- see
563: Fig.6. Note however that in the above simulations with UV field we
564: used a field of the Efstathiou (1992) type which is too hard and
565: intense compared to the more realistic one of Haardt \& Madau (1996)
566: --- see also Sec. 5.3.5. Hence, the suppression in cooling efficiency and
567: the related increase in $L_{X,bol}$ for disc galaxy haloes formed in
568: simulations with a Efstathiou type UV field is
569: somewhat too large.
570:
571: \subsubsection{Effects of gas metallicity}
572: \label{metal}
573: The effects of the metallicity of the gas on the
574: derived $L_{X,bol}$ can be investigated by comparing the squares in
575: Fig.\ref{lxvc} (which have $Z=1/3~Z_{\odot}$) with the rest of the symbols
576: (which have primordial abundance).
577: In the simple cooling flow models, an increase of the metal abundance leads to an
578: increase in the cooling rate and $L_{X,bol}$ (compare the thick and thin
579: curves in Fig.\ref{lxvc}); however this is not what we find from our
580: simulations. The galaxies with $Z=1/3~Z_{\odot}$ have systematically lower $L_{X,bol}$ than the
581: galaxies with primordial abundance, by about a factor of 3-4 for the
582: same $f_b$. This
583: is in agreement with the argument that increasing the cooling
584: efficiency leads to a decrease in $L_{X,bol}$ at $z=0$.
585: Note that for the $Z=1/3~Z_{\odot}$ simulations we used the cooling
586: function of Sutherland \& Dopita (1993) which does not include the
587: effect of a UV field. So we can not completely disentangle the effects
588: of gas metal abundance versus lack of UV field on the X-ray
589: luminosities, but Fig.6 strongly hints that the former is the most important.
590: \begin{figure}
591: \resizebox{\hsize}{!}{{\includegraphics{dlxmhot.ps}}}
592: \caption{Bolometric luminosity relative to the one expected for the simple
593: cooling flow models versus hot gas fraction - see text for details.
594: Symbols as in Fig. 2.}
595: \label{dlxmhot}
596: \end{figure}
597:
598:
599: \subsubsection{Effects of baryon fraction}
600: Comparing in Fig.~2 $L_{X,bol}$ for simulations run with primordial gas and
601: with $f_b$=0.05 and 0.10, respectively, the former are more luminous on average by
602: about a factor of two. Yet again we see how increased cooling efficiency leads
603: to a decreased $L_{X,bol}$ at $z$=0.\\[1cm]
604: Summarizing the above results we find no statistically significant dependence
605: of the X-ray luminosities of the simulated galaxies on numerical resolution.
606: With respect to cooling efficiency there is
607: a general trend of higher cooling efficiency over the
608: course of the simulation to result in less hot gas left in the halo at $z$=0
609: to cool, yielding a lower X-ray luminosity.
610: This is qualitatively demonstrated in Fig. 6, which
611: shows the bolometric luminosity relative to the one expected for the simple
612: cooling flow models versus the hot gas fraction
613: $f_{(hot~gas)}=M_{(hot~gas)}(<r_{vir})/(f_b M_{vir}$), where
614: $M_{(hot~gas)}(<r_{vir})$ is the mass of hot gas (log($T$)$>$4.5) inside of the
615: virial radius $r_{vir}$ and $M_{vir}$ is the total mass (baryonic + DM) inside
616: $r_{vir}$.
617: It is seen that only for hot gas fractions $f_{(hot~gas)}\ga$ 0.4-0.5,
618: requiring a physically implausible parameter combination of
619: $f_b$=0.05, primordial gas and an unrealistically hard and intense UV field, can our
620: models match the $L_{X,bol}$ predicted by the simple cooling flow models.
621:
622: Pulsar dispersion measures can be used to place observational upper
623: limits on the amount of hot gas in the halo of the Milky Way. We find
624: from our simulations that Milky Way sized galaxies formed in
625: primordial gas simulations have about 10$^9 M_{\odot}$ of hot
626: gas inside of
627: 50 kpc at $z$=0 and the ones from the $Z=1/3~Z_{\odot}$ simulations about 10$^8
628: M_{\odot}$. Both values are consistent with the observational upper limits
629: of about $2\cdot10^9M_{\sun}$ from
630: pulsar dispersion measures to the Magellanic Clouds and the globular cluster
631: M53 (Moore \& Davis 1994; Rasmussen 2000).
632:
633: The difference between the $f_b$=0.05 and 0.10 cases is even more pronounced
634: for the 0.2-2 keV X-ray luminosities, as shown in Fig. 3. The reason is that
635: at a {\it given} characteristic circular speed $V_c$ the dynamics of the inner
636: galaxy (where $V_c$ is determined) are more baryon dominated for $f_b$=0.10
637: than for $f_b$=0.05. This in turn means that the hot halo is smaller and
638: cooler for the $f_b$=0.10 case than for the $f_b$=0.05 case.
639: This is demonstrated in Fig. 7, which shows the average temperature of the
640: central hot halo gas (inside of 20 $h^{-1}$kpc) for the 44 simulations. At
641: a given $V_c$ the temperature of the inner, hot halo is systematically shifted
642: to lower values for $f_b$=0.10 as compared to $f_b$=0.05.
643: Hence for the
644: relevant, relatively low temperatures ($T \la 0.3$ keV, or equivalently,
645: $V_c \la$ 300 km s$^{-1}$) less of the emitted radiation has energies above 0.2 keV
646: for the former than for the latter case.
647: On the issue of baryon dominance of the inner galaxy dynamics note that
648: for a {\it
649: given} DM
650: halo, $f_b$=0.10 (as compared to $f_b$=0.05) leads to a larger $V_c$ and a
651: smaller d$v_c(R)$/d$R$ in the outer parts of the disc, where $v_c(R)$
652: is the rotation curve -- this is in line with the findings of Persic, Salucci \& Stel
653: (1996) on the basis of a large observational sample of disc galaxy rotation
654: curves.
655:
656: Finally, as mentioned in Sec.~5.2, the temperature of the Milky
657: Way's inner halo is about 0.15 keV corresponding to
658: 1.5-2$\cdot10^6$~K in agreement with Fig.7.
659: \begin{figure}
660: \resizebox{\hsize}{!}{{\includegraphics{Tx.ps}}}
661: \caption{Average inner halo hot gas temperature versus characteristic circular
662: speed - see text for details. Symbols as in Fig. 2.}
663: \label{Tx}
664: \end{figure}
665:
666:
667: \subsubsection{Effects of dark matter type, cosmology and star formation}
668:
669: We do not find any dependence on the DM type, i.e. whether the
670: simulations are of the WDM or CDM~+~feedback type. Neither do we find any
671: indications of systematic trends with cosmology (SCDM/WDM versus
672: $\Lambda$CDM/WDM) although more overlap in figures 2 and 3 between the two
673: cosmologies would have been desirable (the reason for this lack of
674: overlap is that rather different cosmological volumes were sampled in the two
675: cosmologies --- the box size was $\sim 40 h^{-1}$Mpc for the
676: $\Omega_M$=1 cosmology and $10 h^{-1}$Mpc for the $\Lambda$-cosmology). Note also that in Fig. 6 the results
677: for the different cosmologies fall along the same continuous sequence. The
678: reason why the disc galaxies formed in primordial gas simulations for the
679: $\Omega_m=1$ cosmology tend to have slightly higher relative bolometric
680: luminosity {\it and} $f_{(hot~gas)}$ than the $\Lambda$-cosmology ones is most
681: likely due to the two different models of the meta-galactic UV field used:
682: For the former we used the one suggested by Efstathiou (1992) (see
683: Vedel \etal 1994), whereas for the latter we used the more realistic one
684: from Haardt \& Madau (1996). The former has $z_{reionization} = \infty$ and
685: is considerably harder and, at $z \ga 2$, more intense than the latter
686: which has $z_{reionization} \simeq 6$.
687:
688: Finally, we do not find any dependence on whether star formation is
689: included or not in the simulations.
690:
691:
692:
693:
694:
695: %In semi-analytical models, increasing the baryon fraction is expected
696: %to increase the cooling rate and $L_X$ (compare the thick and thin curves %Fig.\ref{lxvc})
697: %This effect is not observed in our simulations.
698: %Instead galaxies with $f_b$=0.1 and flat cosmology have systematically
699: %smaller $L_X$ than the galaxies with $f_b$=0.05 and open cosmology, in
700: %the region where there is overlap in $v_c$ (again in agreement with
701: %the argument that increasing the cooling efficiency lead to a decrease
702: %in $L_X$).
703: %It is however not possible to determine if the lower $L_X$ of the
704: %galaxies with higher $f_b$ is due to this effect, or due to the fact
705: %that they were simulated in a flat cosmology.
706:
707:
708: \subsection{Mass accretion rates}
709: In Sommer-Larsen \etal (2002) disc gas accretion rates due to cooling-out of
710: hot gas are determined for the $\Lambda$CDM + feedback model disc
711: galaxies considered there (a subset of the sample considered here obtained in simulations with $f_b$=0.10, primordial
712: gas composition and the Haardt \& Madau UV field). For Milky-Way sized
713: galaxies accretion rates in the range 0.3-0.6 $h^{-1}$M$_{\odot}$yr$^{-1}$ ($h$=0.65)
714: are found at $z$=0.
715: One can show that the rate at which hot halo gas cools out is proportional
716: to $L_{{\rm X,bol}} \cdot <\frac{1}{T}>$, where $<\frac{1}{T}>$ is the
717: emissivity
718: weighted inverse temperature of the hot gas. As mentioned previously we find
719: in the current work that $L_{{\rm X,bol}}$ is proportional to about the fifth
720: power of $V_c$, so as $T \propto V_c^2$ we would expect $\dot{M} \propto
721: V_c^3$. This is indeed found by Sommer-Larsen \etal (2002) and is sensible
722: since the $I$-band (and hence approximately mass) Tully-Fisher relation has
723: a logarithmic slope of about three (Giovanelli \etal 1997).
724:
725: Given the trends of $L_X$ with various environmental parameters discussed in
726: section 5.3 we would expect galaxies formed in simulations with $f_b$=0.05
727: and primordial abundance to have about twice as large mass accretion rates,
728: whereas galaxies formed in $Z=1/3~Z_{\odot}$ simulations will have accretion
729: rates 3-4 times lower than the similar primordial gas ones. Hence we expect
730: at $z$=0 a fairly strong trend of the ratio of present to average
731: past accretion rate {\it decreasing} with {\it increasing} cooling efficiency. Indeed
732: Sommer-Larsen \etal (2002) find for their $\Lambda$CDM simulations accretion
733: rates at $z$=1 which are an order of magnitude larger than the ones at $z$=0,
734: so disc galaxies may have been considerably more X-ray luminous in the past
735: than they are today.
736:
737: A detailed analysis of mass accretion rates and high-$z$ X-ray properties of
738: our sample of simulated disc galaxies will be presented in a forthcoming
739: paper.
740:
741:
742: \subsection{Future X-ray observational tests}
743: From the predicted X-ray surface brightness profiles (Fig.4)
744: and the halo temperatures (Fig.7) we have estimated the feasibility for
745: detecting halo emission with XMM-Newton and Chandra using
746: the most recent instrument responses.
747: In order to avoid confusion with X-ray emission originating in the
748: disc we aim at detecting halo emission from nearly edge-on disc
749: galaxies at (vertical) disc heights of 10-15 kpc.
750: Count rates for a 5 kpc high and 40 kpc wide slice (parallel to
751: the disc) at such disc
752: heights were calculated assuming a column density
753: of absorbing neutral hydrogen in the disc of the Milky Way of $n_H=2.5\cdot
754: 10^{20}$~cm$^{-2}$ (corresponding to the typical value for
755: galactic latitudes of $|b|$$\sim$60$^{\circ}$).
756: For galaxies with circular speeds in excess of 300~km~s$^{-1}$ and distances $d$$\la$50 Mpc XMM-Newton, should be able to obtain a $5\sigma$
757: detection at such disc heights in a 10 ksec exposure.
758: Although Chandra has a smaller collecting area than XMM-Newton
759: the Chandra background is generally lower and its superior spatial resolution
760: allows for more efficient removal of contaminating point sources.
761: We thus expect that only slightly longer exposures are required for Chandra detection of halo emission than for XMM-Newton.
762:
763: However, the curve in Fig.4 for the $V_c>300$~km~s$^{-1}$ galaxies represents an
764: optimistic case since the underlying simulations were run with primordial
765: abundances and a strong external UV-field, both increasing the present
766: day X-ray luminosity. As mentioned in sections 5.3.2 and 5.3.3, in more realistic
767: simulations, including metals and a weaker external UV-field,
768: the halo flux is lower by a factor of about 3.
769: In this case, XMM-Newton as well as Chandra should still obtain
770: a $5\sigma$ detection for $V_c>300$~km~s$^{-1}$ galaxies within 25 Mpc in
771: about 25 ksecs.
772:
773: For Milky Way sized galaxies, due to their
774: much lower surface brightness (e.g. Fig.4) and lower halo temperature
775: (the latter making these more sensitive to absorption) the predicted
776: XMM-Newton and Chandra halo count rates are
777: two orders of magnitude lower than for the $V_c>300$~km~s$^{-1}$ galaxies. Detection of
778: X-ray haloes for Milky Way sized galaxies at vertical disc heights of
779: 10-15 kpc will thus have to await
780: future X-ray observatories with much larger collecting areas (Constellation-X
781: and XEUS).
782:
783:
784:
785:
786:
787: \section{Conclusion}
788:
789: We have presented X-ray properties of the hot gas haloes
790: of disc galaxies derived from a large sample of physically realistic
791: gravity/hydro simulations of galaxy formation and evolution.
792:
793: The simulated galaxies follow an L$_{X,bol}$-$V_c$ relation
794: with approximately the same slope as expected from simple cooling flow
795: models ($L_{X,bol}\propto V_c^5$), but shifted to lower
796: $L_{X,bol}$, and with a significant
797: scatter (approximately a 50\% rms dispersion for a given choice of
798: physical parameters).
799:
800: The total bolometric X-ray luminosities of the disc galaxy haloes are
801: up to two orders of magnitude less than predicted by simple cooling
802: flow models.
803: Hence, contrary to the order of magnitude discrepancy between simple cooling flow models and observations
804: found by Benson \etal (2000),
805: our models are in agreement with ROSAT observations of
806: the three massive highly inclined spirals NGC 2841, NGC 4594 and NGC
807: 5529.
808: Furthermore, we find that our models are
809: consistent with recent estimates of the diffuse 0.2-2 keV X-ray
810: luminosity of the Milky Way, and also that the amount of hot gas in
811: the haloes of our simulated, Milky Way sized disc galaxies is consistent with upper limits from pulsar
812: dispersion measures toward the Magellanic Clouds and the globular
813: cluster M53.
814:
815: In contrast to what is predicted by simple cooling flow models, we
816: find that {\it increasing} the cooling efficiency of the halo gas leads to
817: a {\it decrease} in the present day $L_X$.
818: The reason for this is that increasing the cooling efficiency over
819: the course of a simulation results in less hot gas in the halo at $z$=0
820: to cool (because the total amount of gas available at any given time
821: is always limited to the gas inside of the virial radius). This in turn leads to lower present day accretion rates and
822: lower $L_{X,bol}$.
823: The two most important physical parameters controlling the
824: X-ray luminosities are the baryon fraction and the gas abundance:
825: for a {\it given} characteristic circular speed $V_c$,
826: increasing the baryon fraction from $f_b$=0.05 to $f_b$=0.1
827: decreases $L_{X,bol}$ by a factor of about two, and similarly
828: increasing the gas abundance from primordial to $Z=1/3~Z_{\odot}$
829: results in a decrease in $L_{X,bol}$ of a factor 3-4.
830:
831: Concerning the spatial distribution of X-ray emission in the hot gas
832: haloes we find that this is centrally concentrated: about 95\% of the
833: emission originates within the inner $r\la$20 kpc.
834:
835: For their $\Lambda$CDM, $f_b$=0.1 and primordial gas abundance simulations Sommer-Larsen \etal (2002) find present
836: day mass accretion rates of 0.5-1 $M_{\sun}$yr$^{-1}$ for Milky Way
837: sized disc galaxies, compatible with observational limits (J. Silk,
838: private communication). As the mass accretion rate effectively is
839: proportional to $L_{X,bol}$ one would expect even lower accretion
840: rates for the more realistic case of some level of metal enrichment of
841: the hot gas. Furthermore, Sommer-Larsen \etal (2002) find mass accretion rates
842: which are an order of magnitude larger at redshift $z\sim$1. Hence it is quite
843: likely that disc galaxies were considerably X-ray brighter in the
844: past.
845:
846:
847: Forthcoming XMM-Newton and Chandra observations of massive, nearby, edge-on
848: disc galaxies will provide constraints on the present
849: models, in particular the main physical parameters: the baryon
850: fraction and the hot gas abundance. Future observations to redshifts
851: $z\ga$1 may be used to constrain the models further.
852:
853:
854:
855:
856: % ($\dot{M}\sim 0.5M_{\sun}yr^{-1}$ and $L_{X,bol}\sim 10^{40}
857: %$erg s$^{-1}$ respectively for a MW sized galaxy).
858:
859: %The X-ray emission is found to be very concentrated. More than 95\% of the
860: %emission originates within 10 kpc of the disc. (***** KP: XMM, CHANDRA feasibility*****)
861:
862:
863:
864:
865:
866:
867:
868:
869:
870:
871:
872:
873:
874:
875:
876:
877:
878:
879:
880:
881:
882: \section*{Acknowledgments}
883: This project was supported by the Danish Natural Science Research
884: Council (SNF) and by Danmarks Grundforskningsfond through its support
885: for the establishment of the Theoretical Astrophysics Center.
886: We thank R. Bower, M. G\"otz, B. Moore, L. Portinari and J. Silk for useful discussions.
887:
888:
889: \begin{thebibliography}{99}
890:
891:
892: %\bibitem[\protect\citename{White \& Frenk} 1991]{whitefrenk} White,
893: %S. D. M., Frenk, C. S. 1991, ApJ, 379, 52
894:
895: \bibitem[\protect\citename{Benson et al.} 2000]{benson} Benson, A. J.,
896: Bower, R. G., Frenk, C.S., White, S. D. M., 2000, MNRAS, 314, 557
897:
898: \bibitem[\protect\citename{Efstathiou} 1992]{efstathiou}
899: Efstathiou, G., 1992, MNRAS, 256, P43
900:
901: \bibitem[\protect\citename{Ettori \& Fabian} 1999]{EF}
902: Ettori, S., Fabian, A.C., 1999, MNRAS, 305, 834
903:
904: \bibitem[\protect\citename{Giovanelli et al.} 1997]{G.97}
905: Giovanelli, R., \etal 1997, ApJ, 477, L1
906:
907: \bibitem[Georgantopoulus et al.\ 1996]{geor1996} Georgantopoulus, I.,
908: Stewart, G.C., Shanks, T., Boyle, B.J., Griffiths, R.E. 1996,
909: MNRAS, 280, 276
910:
911: \bibitem[\protect\citename{Haardt \& Madau} 1996]{HM}
912: Haardt, F., Madau, P., 1996, ApJ, 461, 20
913:
914: \bibitem[\protect\citename{Monaghan \& Lattanzio} 1985]{monagan}
915: Monaghan, J.J., Lattanzio, J.C., 1985, A\&A, 149, 135
916:
917: \bibitem[\protect\citename{Moore \& Davis} 1994]{moore}
918: Moore, B., Davis, M, 1994, MNRAS, 270, 209
919:
920: \bibitem[\protect\citename{Mewe et al.} 1986]{mewe}
921: Mewe, R., Lemen, J.R., van den Oord, G.H.J, 1986, A\&A, 65, 511
922:
923: \bibitem[\protect\citename{Navarro \& White} 1994]{NW94}
924: Navarro, J. F., White, S. D. M., 1994, MNRAS, 267, 401
925:
926: \bibitem[\protect\citename{Navarro, Frenk \& White} 1996]{NFW}
927: Navarro, J. F., Frenk C. S., White, S. D. M., 1996, ApJ, 462, 563
928:
929: \bibitem[\protect\citename{Pagel} 1997]{pagel}
930: Pagel, B. E. J., 1997, Nucleosynthesis and Chemical Evolution of
931: Galaxies (Cambridge: Cambridge Univ. Press)
932:
933: \bibitem[Parmar et al.\ 1999]{parm1999} Parmar, A.N., Guainazzi, M.,
934: Oosterbroek, T., Orr, A., Favata, F., Lumb, D., Malizia, A. 1999,
935: A \& A, 345, 611
936:
937: \bibitem[\protect\citename{Percival et al.} 2001]{per}
938: Percival, W.J. et al., 2001, MNRAS, 327, 1297
939:
940: \bibitem[\protect\citename{Persic et al.} 1996]{PSS96}
941: Persic, M., Salucci, P., Stel, F., 1996, MNRAS, 281, 27
942:
943: \bibitem[\protect\citename{Pietsch et al.} 1997]{pietsch}
944: Pietsch, W. et al., 2001, MNRAS, 365, L174
945:
946: \bibitem[\protect\citename{Pietz et al.} 1998]{pietz}
947: Pietz, J. et al., 1998, A\&A, 332, 55
948:
949: \bibitem[\protect\citename{Rasmussen} 2000]{Rasmussen}
950: Rasmussen, J., 2000, Master Thesis, University of Copenhagen
951:
952: \bibitem[\protect\citename{Read, Ponman \& Strickland } 1997]{read}
953: Read, A.M., Ponman, T.J., Strickland, D. K., 1997, MNRAS, 286, 626
954:
955: \bibitem[\protect\citename{Rocha-Pinto \& Marciel}1996]{rocha-pinto}
956: Rocha-Pinto, H. J., Maciel, W. J., 1996,
957: MNRAS, 273, 447
958:
959: \bibitem[\protect\citename{Sommer-Larsen \& Dolgov} 2001]{sommerlarsendolgov}
960: Sommer-Larsen, J., Dolgov, A., 2001, ApJ, 551, 608
961:
962: \bibitem[\protect\citename{Sommer-Larsen et al.} 1999]{sommerlarsen99}
963: Sommer-Larsen, J., Gelato, S., Vedel, H., 1999, ApJ, 519, 501
964:
965: \bibitem[\protect\citename{Sommer-Larsen et al.} 2002]{sommerlarsen02}
966: Sommer-Larsen, J., G\"otz, M., Portinari, L., 2002, in preparation
967:
968: \bibitem[\protect\citename{Strickland \& Stevens} 2000]{strickland}
969: Strickland, D. K., Stevens, I. R., 2000, MNRAS, 314, 511
970:
971: \bibitem[\protect\citename{Sutherland \& Dopita} 1993]{sd93}
972: Sutherland, R. S., Dopita, M. A., 1993, ApJS, 88, 253
973:
974: \bibitem[\protect\citename{Vedel, Hellsten and Sommer-Larsen} 1994]{vedel}
975: Vedel, H., Hellsten, U., Sommer-Larsen, J. 1994, MNRAS, 271, 743
976:
977: \bibitem[\protect\citename{Wang} 1997]{wang97}
978: Wang, Q., 1997, astro-ph/9710144
979:
980: \bibitem[\protect\citename{Wang et al.} 2001]{wang}
981: Wang, Q. et al., 2001, ApJ, 555, L99
982:
983:
984: \end{thebibliography}
985:
986:
987:
988: \label{lastpage}
989:
990: \end{document}
991:
992:
993:
994:
995:
996:
997:
998: \begin{figure}
999: \resizebox{\hsize}{!}{{\includegraphics{lxappplot_dist13.8000_paper.ps}}}
1000: \caption{Bolometric luminosity in the annulus considered by \cite{benson}}
1001: \label{lxappdist}
1002: \end{figure}
1003: (a ring around the galaxy, with an
1004: inner radius of 5 arcmin and an outer radius of 18 arcmin) of all
1005: the galaxies in our simulated sample (assuming a distance of 13.8 Mpc).
1006: All galaxies are viewed ``edge on'', the error bars represents the
1007: difference in $L_X$ resulting from rotating all the galaxies 90{\degr}
1008: about their spin axis before applying the annulus. The observed data point
1009: (indicated by a cross) and its error bars are adopted from Benson \etal
1010: \shortcite{benson}}
1011: There is little overlap in $V_c$ between galaxies simulated in a flat
1012: $\Omega_M=1$ cosmology (small symbols in
1013: Fig. \ref{lxvc}) and galaxies simulated in a flat $\Omega_{\Lambda}+
1014: \Omega_M=1$ cosmology (large symbols in Fig. \ref{lxvc}).
1015: Most of the galaxies simulated in the flat cosmology have baryon
1016: fraction $f_b$=0.1 (filled symbols), while most of the galaxies
1017: simulated in the open cosmology have baryon fraction $f_b$=0.05 (open
1018: symbols).
1019: This makes it tricky to single out effects of baryon
1020: fraction and cosmology in the simulations.
1021:
1022:
1023:
1024:
1025:
1026:
1027:
1028:
1029:
1030:
1031:
1032:
1033:
1034:
1035:
1036:
1037: