1: \documentstyle[sprocl]{article}
2:
3: \font\eightrm=cmr8
4:
5: \input{psfig}
6:
7: \bibliographystyle{unsrt} %for BibTeX - sorted numerical labels by
8: %order of first citation.
9: \arraycolsep1.5pt
10:
11: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
12: %%BEGINNING OF TEXT
13: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
14:
15: \begin{document}
16:
17: \title{THE MORPHOLOGICAL EVOLUTION OF MERGER REMNANTS}
18:
19: \author{BARBARA RYDEN \& JEREMY TINKER}
20:
21: \address{The Ohio State University, Dept. of Astronomy,\\
22: 140 W. 18th Avenue, Columbus OH 43210\\
23: E-mail: ryden, tinker@astronomy.ohio-state.edu}
24:
25: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
26: % You may repeat \author \address as often as necessary %
27: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
28:
29: \maketitle\abstracts{ Elliptical galaxies formed in a major
30: merger have a tendency to become more nearly spherical
31: with time, thanks to the gravitational effect of their
32: central black hole (or black holes). Observational results
33: indicate that elliptical galaxies with older stellar
34: populations ($t > 7.5 {\rm\,Gyr}$) have rounder central
35: isophotes than ellipticals with younger stellar populations.
36: In addition, the older ellipticals tend to have core profiles,
37: while the younger ellipticals have power-law profiles.
38: Numerical simulations of galaxy mergers indicate that
39: if one or both of the progenitors have a central black
40: hole with mass $\sim 0.2$\% of the stellar mass, then
41: the effect of the black hole(s) is to make the central
42: regions of the remnant rounder, with a characteristic
43: time scale of a few gigayears.
44: }
45:
46: \section{Analytic}
47: In a hierarchical clustering scenario, elliptical galaxies
48: form by the merger of smaller stellar systems. Mergers of
49: equal-mass progenitors tend to form fairly flattened
50: systems; after violent relaxation, the ratio of the
51: shortest to longest axis of the merger remnant is
52: typically $c/a \sim 0.5$ \cite{ba92,sp00}. However,
53: once an elliptical merger remnant has completed violent
54: relaxation, its shape does not remain constant.
55: In the central regions of the galaxy -- well
56: inside the effective radius -- the morphological
57: evolution is driven by two-body relaxation, the
58: result of close gravitational encounters between
59: the point masses of which the galaxy is made. The
60: net effect of two-body relaxation is to make a galaxy
61: more nearly spherical with time.
62:
63: Consider an idealized case of two-body relaxation in
64: which a mass $M$ is plunked down in an isothermal
65: stellar system with velocity dispersion $\sigma$.
66: The introduced mass will disrupt the orbits of stars
67: which come within a critical distance $b \sim G M / \sigma^2$.
68: If the mass $M$ is just another star, with mass $M \sim 1
69: {\rm\,M}_\odot$, then stars will have to come within
70: a distance $b \sim 5 {\rm\,R}_\odot ( \sigma / 200 {\rm\,km}
71: {\rm\,s}^{-1} )^{-2}$ before their orbits are randomized.
72: Thus, for a typical elliptical galaxy, stars must come
73: within a few stellar radii of each other for two-body
74: relaxation to occur, and the star/star relaxation time
75: is much longer than a Hubble time. If stars were the
76: only point masses which elliptical galaxies contained, we
77: would thus conclude that the effects of two-body relaxation
78: on the structure of ellipticals are negligibly small
79: so far. However, there's more to a galaxy than stars.
80: Most, if not all, elliptical galaxies contain central
81: black holes, with a mass given by the relation\cite{fm00}
82: $M_{\rm BH} \approx 10^8 {\rm\,M}_\odot ( \sigma / 200
83: {\rm\,km} {\rm\,s}^{-1} )^{4.8}$. With this black hole mass,
84: a star coming within a distance $b \sim 10 {\rm\,pc}
85: ( \sigma / 200 {\rm\,km} {\rm\,s}^{-1} )^{2.8}$ will
86: have its orbit disrupted. Thus,
87: relaxation due to star/black-hole encounters will be
88: vastly more effective than relaxation due to star/star
89: encounters. The net effect of the central black hole
90: will be to increase the entropy of the stellar system
91: and to make it more nearly spherical with time.
92:
93: It should also be noted that if an elliptical forms
94: by the merger of two progenitors, each with a central
95: black hole, a binary black hole may exist for many
96: gigayears before dynamical friction, gas dynamical
97: effects, and gravitational radiation will cause the
98: two black holes to coalesce. Three-body interactions
99: between a star and a bound black hole binary will
100: generally increase the star's kinetic energy. Thus,
101: binary black holes have been proposed as a mechanism
102: for lowering the stellar density in the central regions
103: of an elliptical and creating a `core' profile \cite{em91,fa97}.
104:
105: \section{Observational}
106: Given the brevity of human life, we cannot sit
107: and watch for a few billion years while a post-merger
108: elliptical becomes rounder with time; nor can we
109: take the time to circumnavigate a galaxy and discover
110: its true three-dimensional shape at a given time.
111: The best we can do, to test our belief that merger
112: remnants become rounder with time, is to examine
113: a sample of elliptical galaxies and see whether
114: the apparent shape of a galaxy is correlated with
115: the time elapsed since it last underwent a major
116: merger.
117:
118: Estimating the time that has passed since an elliptical
119: galaxy's last major merger is not a simple or straightforward
120: task. However, if the merger in question involved a
121: pair of reasonably gas-rich galaxies, then the merger
122: will be accompanied by a burst of star formation that
123: will leave its spectroscopic mark on the galaxy.
124: Terlevich and Forbes have recently compiled a catalog\cite{tf01}
125: of spectroscopic galaxy ages, based on a homogeneous
126: data set of galaxies with high-quality H$\beta$ and
127: [MgFe] absorption line indices. The stellar population
128: model of Worthey\cite{wo94} is used to assign an age
129: to the stellar population of each galaxy in the catalog.
130: For the 74 elliptical galaxies in the Terlevich \&
131: Forbes catalog, we searched the published literature
132: for isophotal fits, and found the apparent axis ratio
133: $q \equiv b/a$ at six reference radii: $R \equiv (ab)^{1/2} =
134: R_e /16$, $R_e/8$, $R_e/4$, $R_e/2$, $R_e$, and $2 R_e$.
135: Details of the analysis, for those who love details,
136: are given by Ryden, Forbes, \& Terlevich\cite{rf01}.
137:
138: \begin{figure}
139: \psfig{figure=rt_fig1.ps,height=6.0in}
140: \caption{Isophotal axis ratio $q$ versus estimated age $t$
141: of the central stellar population. Axis ratios are measured
142: at $R = R_e/16$, $R_e/8$, $R_e/4$, $R_e/2$, $R_e$, and
143: $2 R_e$. Galaxies with core profiles are indicated by
144: squares, galaxies with power-law profiles are indicated
145: by triangles, and galaxies with unknown profile type
146: are indicated by open circles.}
147: \label{fig:forbes}
148: \end{figure}
149:
150: Figure 1 is a plot of $q$ versus the spectroscopic age $t$
151: at the six reference radii. Particularly at the innermost
152: radius, $R = R_e/16$, there is a correlation between $q$ and
153: $t$, with `old' ellipticals tending to be
154: rounder than `young' ellipticals.
155: A Kolmogorov-Smirnov test, comparing the distribution of $q(R_e/6)$
156: for galaxies with $t \leq 7.5 {\rm\,Gyr}$ to the distribution
157: for galaxies with $t > 7.5 {\rm\,Gyr}$, reveals that the
158: distributions differ significantly, with $P_{\rm KS} = 0.00034$.
159:
160: The question `Why are there so many round old ellipticals?'
161: is similar to the question `Why are there so many little old
162: ladies?' It is tempting to interpret the prevalence of little
163: old ladies as being due purely to the evolution of individuals, with
164: ladies tending to become littler as they grow older. However,
165: there are other effects at work as well. For instance, extremely
166: large ladies tend to die prematurely; in addition, today's population
167: of old ladies grew up when nutritional standards were lower, thus
168: resulting in reduced adult stature. Both these effects, neither of
169: which involves the morphological evolution of individual ladies,
170: contribute to the predominance of little old ladies over large
171: old ladies. Similarly, the predominance over round old ellipticals
172: over flattened old ellipticals is not necessarily due to the morphological
173: evolution of individual galaxies.
174:
175: The difference in apparent shape between galaxies with young stellar
176: populations and those with old stellar populations is tied, in a
177: most intriguing manner, to the core/power-law distinction. Elliptical
178: galaxies with power-law profiles have luminosity densities which are
179: well fit be a pure power law all the way to the limit of resolution;
180: ellipticals with core profiles, by contrast, have densities which
181: show a break to a shallower inner slope\cite{fe94,fo95,la95}.
182: In Figure 1, core ellipticals are designated by squares, power-law
183: ellipticals are designated by triangles, and ellipticals of unknown
184: profile type are designated by empty circles. {\it Note that the `old'
185: ellipticals tend to have core profiles and round central isophotes,
186: while the `young' ellipticals tend to have power-law profiles
187: and flattened central isophotes.} The 29 known core galaxies in
188: our sample have a mean and standard deviation for their estimated
189: stellar ages of $t = 8.6 \pm 3.3 {\rm\,Gyr}$; the 22 known power-law
190: galaxies have $t = 6.9 \pm 3.5 {\rm\,Gyr}$.
191:
192: \section{Numerical}
193:
194: The observational results are consistent with a scenario in
195: which elliptical galaxies are formed in a major merger,
196: then evolve to become more nearly spherical with time.
197: However, they do not compel such a scenario -- remember
198: the cautionary tale of the little old ladies! Fortunately,
199: numerical simulations of galaxy mergers, and of the evolution
200: of merger remnants, can be run on timescales much shorter than
201: a gigayear (and, more to the point, shorter than than the lifetime of
202: a graduate student.)
203:
204: We ran n-body simulations (with no attempt to include
205: gas dynamical effects) of the merger of a pair of
206: disk/bulge/halo galaxies.
207: In our merger simulations, each progenitor has a
208: disk:bulge:halo mass ratio of 1:1:5.8. (The
209: progenitors can be thought of as a pair of S0 galaxies,
210: with big bulges and no gas). The disk, the bulge,
211: and the halo each contain 16K particles. We ran
212: three different merger simulations, differing only
213: in the mass of the central black hole assigned
214: to each progenitor galaxy. One simulation contained
215: no central black holes. In the next simulation, the
216: progenitor galaxies contained central black holes equal
217: in mass to 0.2\% of the total stellar mass (disk + bulge).
218: In the final simulation, the progenitors contained black
219: holes equal in mass to 2\% of the total stellar mass.
220:
221: \begin{figure}[t]
222: \psfig{figure=rt_fig2.ps,height=3.65in}
223: \caption{The top panel shows the evolution of the
224: intermediate-to-long axis ratio (b/a) of a numerical
225: merger remnant. THe bottom panel shows the evolution
226: of hte short-to-long axis ratio (c/a) for the same
227: remnant. In each panel, the heavy solid line indicates
228: the 5\% of the particles which are most tightly bound.
229: In this simulation, the merging progenitors did NOT
230: contain central black holes.}
231: \label{fig:nobh}
232: \end{figure}
233:
234: We used the n-body code GADGET\cite{syw01} for all
235: integrations. GADGET is a tree code designed to run on distributed
236: memory, multi-processor computers. It employs continuously variable
237: timesteps which are individual to each particle. The timesteps are
238: computed with an accuracy parameter,
239: $\eta$, set to $0.02$.
240: The gravitational smoothing lengths for the disk, bulge, and halo
241: particles were 0.08, 0.08, and 0.4 respectively. For force calculations
242: between any particle and a black hole, a smoothing length of 0.001 was
243: used. (For reference, the disk scale length of the progenitor galaxies
244: was 1.0.)
245:
246: In the first simulation, whose results are presented in
247: Figure 2, the merging galaxies contained no central black
248: holes. Note that both the intermediate-to-long axis ratio
249: (illustrated in the upper panel of Figure 2) and the
250: short-to-long axis ratio (illustrated in the lower panel)
251: evolve steadily toward unity in this simulation, despite
252: the absence of a central black hole. This is a spurious
253: two-body relaxation effect, resulting from the coarseness
254: of our simulation; instead of being made of tens of billions
255: of stars, the `luminous' portions of our simulated galaxies
256: contain only tens of thousands of mass points.
257:
258:
259: \begin{figure}[t]
260: \psfig{figure=rt_fig3.ps,height=3.65in}
261: \caption{As in Figure 2, but for a simulation
262: in which the merging progenitors have moderately
263: massive central black holes (equal in mass to
264: 0.2\% of the total stellar mass).}
265: \label{fig:bh002}
266: \end{figure}
267:
268: In Figure 3, showing the evolution in shape of a merger
269: remnant with moderate mass black holes (0.2\% of the
270: total stellar mass of the progenitors), we see that the
271: evolution toward a spherical shape is more rapid than
272: in the absence of black holes. Moreover, the drive toward
273: a spherical shape is most rapid for the 5\% most tightly
274: bound particles (the heavy solid line in Figure 3) than
275: for the 50\% most tightly bound (the light solid line).
276: In short, the added black holes drive the central, most
277: tightly bound, regions of the merger remnant toward
278: a spherical shape on gigayear timescales.
279:
280: \begin{figure}[t]
281: \psfig{figure=rt_fig4.ps,height=3.65in}
282: \caption{As in Figure 2, but for a simulation
283: in which the merging progenitors have extremely
284: massive central black holes (equal in mass to
285: 2\% of the total stellar mass).}
286: \label{fig:bh02}
287: \end{figure}
288:
289: Adding black holes an order of magnitude more massive,
290: as shown in Figure 4, dramatically shortens the time
291: for making the merger remnant spherical. With big black
292: holes, equal to 2\% of the total stellar mass, the
293: merger remnant rapidly becomes very nearly oblate (that
294: is, the ratio b/a, shown in the upper panel of Figure 4,
295: rapidly approaches one.) The ratio c/a, shown in the
296: lower panel, approaches unity more gradually. However,
297: we can conclude that if merging galaxies contained extremely
298: massive black holes, equal to one or two percent of their
299: total stellar mass, then merger remnants would become
300: nearly oblate on timescales shorter than a gigayear.
301: The relative scarcity of nearly circular isophotes ($q
302: > 0.95$) in the central regions of elliptical galaxies --
303: see the upper left panel of Figure 1 -- argues that
304: elliptical galaxies are probably not oblate in their
305: central regions. However, given the relatively small
306: size of the Terlevich-Forbes sample, this is not a
307: chiseled-in-stone conclusion.
308:
309: More data are needed (unsurprisingly). A larger sample size
310: of observational data will help to pin down the relationship
311: among galaxy age, isophote shape, and luminosity profile type. Much
312: work remains to be done (also unsurprisingly). Higher-resolution
313: numerical simulations will reduce the ugly effects of
314: spurious two-body relaxation and will enable us to focus
315: on the physically real effects of massive black holes.
316:
317: \section*{References}
318: \begin{thebibliography}{99}
319:
320: \bibitem{ba92} Barnes, J. 1992, ApJ, 393, 484
321: \bibitem{em91} Ebisuzaki, T., Makino, J., \& Okumura, A. K. 1991,
322: Nature, 354, 212
323: \bibitem{fa97} Faber, S. M., et al. 1997, AJ, 114, 1771
324: \bibitem{fm00} Ferrarese, L., \& Merritt, D. 2000, ApJ, 539, L9
325: \bibitem{fe94} Ferrarese, L., van den Bosch, F. C., Ford, H. C.,
326: Jaffe, W., \& O'Connell, R. W. 1994, AJ, 108, 1598
327: \bibitem{fo95} Forbes, D. A., Franx, M., \& Illingworth, G. D. 1995,
328: AJ, 109, 1988
329: \bibitem{la95} Lauer, T. R., et al. 1995, AJ, 110, 2622
330: \bibitem{rf01} Ryden, B. S., Forbes, D. A., \& Terlevich, A. I.
331: 2001, MNRAS, in press
332: \bibitem{sp00} Springel, V. 2000, MNRAS, 312, 859
333: \bibitem{syw01} Springel, V., Yoshida, N., \& White, S.D.M, 2001, New
334: Astron., 6, 79
335: \bibitem{tf01} Terlevich, A. I., \& Forbes, D. A. 2001, MNRAS, submitted
336: \bibitem{wo94} Worthey, G. 1994, ApJS, 95, 107
337:
338: \end{thebibliography}
339:
340: \end{document}
341:
342:
343:
344: