1: %%
2: %% Beginning of file 'sample.tex'
3: %%
4: %% Modified 2005 June 21
5: %%
6: %% This is a sample manuscript marked up using the
7: %% AASTeX v5.x LaTeX 2e macros.
8:
9: %% The first piece of markup in an AASTeX v5.x document
10: %% is the \documentclass command. LaTeX will ignore
11: %% any data that comes before this command.
12:
13: %% The command below calls the preprint style
14: %% which will produce a one-column, single-spaced document.
15: %% Examples of commands for other substyles follow. Use
16: %% whichever is most appropriate for your purposes.
17: %%
18: %%\documentclass[12pt,preprint]{aastex}
19:
20: %% manuscript produces a one-column, double-spaced document:
21:
22: \documentclass[useAMS,usenatbib]{mn2e}
23:
24: %% preprint2 produces a double-column, single-spaced document:
25:
26: %% \documentclass[preprint2]{aastex}
27:
28: %% Sometimes a paper's abstract is too long to fit on the
29: %% title page in preprint2 mode. When that is the case,
30: %% use the longabstract style option.
31:
32: %% \documentclass[preprint2,longabstract]{aastex}
33:
34: %% If you want to create your own macros, you can do so
35: %% using \newcommand. Your macros should appear before
36: %% the \begin{document} command.
37: %%
38: %% If you are submitting to a journal that translates manuscripts
39: %% into SGML, you need to follow certain guidelines when preparing
40: %% your macros. See the AASTeX v5.x Author Guide
41: %% for information.
42:
43: \usepackage{epsfig}
44: \usepackage{myaasmacros}
45:
46: \newcommand{\vdag}{(v)^\dagger}
47: \newcommand{\myemail}{jesseit@usm.uni-muenchen.de}
48:
49: %% You can insert a short comment on the title page using the command below.
50:
51:
52:
53: %% If you wish, you may supply running head information, although
54: %% this information may be modified by the editorial offices.
55: %% The left head contains a list of authors,
56: %% usually a maximum of three (otherwise use et al.). The right
57: %% head is a modified title of up to roughly 44 characters.
58: %% Running heads will not print in the manuscript style.
59:
60:
61: %% This is the end of the preamble. Indicate the beginning of the
62: %% paper itself with \begin{document}.
63:
64: \begin{document}
65:
66: %% LaTeX will automatically break titles if they run longer than
67: %% one line. However, you may use \\ to force a line break if
68: %% you desire.
69:
70: \title[Specific Angular Momentum of Disc Merger Remnants and the $\lambda_R$-Parameter]{Specific Angular
71: Momentum of Disc Merger Remnants and the $\lambda_R$-Parameter}
72:
73: %% Use \author, \affil, and the \and command to format
74: %% author and affiliation information.
75: %% Note that \email has replaced the old \authoremail command
76: %% from AASTeX v4.0. You can use \email to mark an email address
77: %% anywhere in the paper, not just in the front matter.
78: %% As in the title, use \\ to force line breaks.
79:
80: \author[R. Jesseit, M. Cappellari, T. Naab, E. Emsellem and A. Burkert]
81: {Roland Jesseit $^{1}$\thanks{E-mail: jesseit@usm.uni-muenchen.de}, Michele Cappellari$^{2}$, Thorsten
82: Naab$^{1}$, Eric Emsellem$^{3}$ \newauthor and Andreas Burkert$^{1}$\\
83: $^1$ Universit\"atssternwarte M\"unchen, Scheinerstr 1, 81679 M\"unchen, Germany \\
84: $^2$ Sub-Department of Astrophysics, University of Oxford, Denys Wilkinson
85: Building, Keble Road, Oxford, OX1 3RH, UK\\
86: $^3$ Universit\'e de Lyon, Lyon, F-69003, France ; Universit\'e Lyon~1,
87: Observatoire de Lyon, 9 avenue Charles Andr\'e, Saint-Genis
88: Laval, \\ F-69230, France ;
89: CNRS, UMR 5574, Centre de Recherche Astrophysique
90: de Lyon ; Ecole Normale Sup\'erieure de Lyon,\\ Lyon, F-69007, France\\}
91: \date{Draft version. Accepted ??ccn?. Received ??? in original form ???}
92: \pagerange{\pageref{firstpage}--\pageref{lastpage}} \pubyear{2002}
93: \maketitle
94: \label{firstpage}
95:
96:
97: \begin{abstract}
98: We use two-dimensional kinematic maps of simulated binary disc mergers to investigate
99: the $\lambda_R$-parameter, which is a luminosity weighted measure
100: of projected angular momentum per unit mass. This parameter was introduced to subdivide the
101: SAURON sample of early type galaxies in so called fast $\lambda_R > 0.1$ and slow
102: rotators $\lambda_R < 0.1$. Tests on merger remnants reveal that $\lambda_R$ is a robust indicator
103: of the true angular momentum content in elliptical galaxies. We find the same range
104: of $\lambda_R$ values in our merger remnants as in the SAURON galaxies.
105: The merger mass ratio is decisive in creating a slow or a fast rotator in a single binary merger, the
106: former being created mostly in an equal mass merger. Slow rotators have a $\lambda_R$ which does not
107: vary with projection. The confusion rate with face-on fast rotators is very small. Merger with low
108: gas fractions form slow rotators with smaller ellipticities and are in much better agreement with the
109: SAURON slow rotators. Remergers of merger remnants are slow rotators but tend to have too high
110: ellipticities. Fast rotators maintain the angular momentum content from the progenitor disc galaxy
111: if merger mass ratio is high. Some SAURON galaxies have values of $\lambda_R$ as high as our
112: progenitor disc galaxies.
113:
114: \end{abstract}
115:
116: %% Keywords should appear after the \end{abstract} command. The uncommented
117: %% example has been keyed in ApJ style. See the instructions to authors
118: %% for the journal to which you are submitting your paper to determine
119: %% what keyword punctuation is appropriate.
120: \begin{keywords}
121: methods: analytical -- methods: N-body simulations -- galaxies:
122: elliptical and lenticular, cD -- galaxies: formation -- galaxies:
123: evolution -- galaxies: fundamental parameters -- galaxies: kinematics and dynamics
124: \end{keywords}
125:
126: \section{Introduction}
127: The connection between projected and the intrinsic properties of elliptical galaxies is a notorious problem.
128: The Hubble classification of early-type galaxies is based on their elliptical shape, which was the natural observable
129: distinctive feature. It was already pointed out by Hubble that this sequence is only a sequence of apparent shape
130: depending on the inclination under which we observe a galaxy. When kinematic observations of ellipticals became available,
131: they showed that some of them rotate too slowly to be shaped by
132: rotation alone \citep{1977ApJ...218L..43I}.
133: \citet{1978MNRAS.183..501B} proposed that these systems are probably shaped to a large
134: extend by anisotropic pressure. This is best seen when the apparent
135: flattening is plotted versus the balance between ordered and unordered motion $(v/\sigma, \epsilon)$ using the relation of
136: these observables for an isotropic rotator as a reference point. Observations also showed that the amount of
137: rotation varies with luminosity
138: \citep{1983ApJ...266...41D} and isophotal shape \citep{1988A&A...193L...7B}. Isophotal shapes as measured by the $a_4$ parameter
139: can either appear boxy ($a_4 <0$) or discy ($a_4 >0$). Boxy and discy galaxies seem to form a true dichotomy as the isophotal shape
140: correlates with on first sight unconnected properties like X-ray luminosity \citep{1989A&A...217...35B} and central surface
141: density slope (\citealp{1994AJ....108.1598F};\citealp{1995AJ....110.2622L};\citealp{1997AJ....114.1771F}). This led
142: \citet{1996ApJ...464L.119K} to propose a revision of the Hubble sequence using the isophotal shape to cast elliptical
143: galaxies into boxy and discy ellipticals, which might be a classification more closely connected to their formation
144: history than apparent ellipticities \citep{2003ApJ...597..893N}.\\
145:
146: Kinematic properties are normally measured from long-slit data at the major axis ($v$) and a
147: central aperture ($\sigma$). But these are very special positions, which might or might not be representative
148: for the overall structure of the galaxy. It is now possible to extract the line-of-sight velocity distributions
149: (LOSVDs) from the full two-dimensional field of a galaxy with modern integral-field instruments. The SAURON
150: survey pioneered the first comprehensive sample of nearby E/S0 galaxies \citep{2002MNRAS.329..513D} with two-dimensional
151: kinematic data. Some of these galaxies exhibit very complex velocity or velocity dispersion fields which are not
152: easily captured by the usual $v/\sigma,\epsilon$ parameterization \citep{2004MNRAS.352..721E}. It can happen
153: that two galaxies with similar long-slit measurements of ellipticity and $v/\sigma$ have morphologically
154: different kinematic fields, e.g. one galaxy has a kinematically decoupled component and the other a regular
155: velocity field. \citet{2007MNRAS.379..401E}, henceforth EM07, therefore introduced a luminosity weighted
156: measure of the specific line-of-sight angular momentum
157: \begin{equation}
158: \lambda_R=\frac{<R|V|>}{<R\sqrt{V^2+\sigma^2}>},
159: \end{equation}
160: where the pointed brackets are luminosity averages. $\lambda_R$ is designed to account for radial changes of angular
161: momentum content in kinematic fields. This is different to the measure $<V>/<\sigma>$ proposed by \citet{2005MNRAS.363..937B}
162: to exploit integral field kinematic data. However, \citet{2005MNRAS.363..937B} also stressed that with
163: integral field data available, the connection between internal and projected structure of a galaxy is much less ambiguous.
164:
165: The E/S0 galaxies are cast into the category of slow rotator or fast rotator according to their measured
166: $\lambda_R$ (EM07): slow rotators have a $\lambda_R < 0.1$ and fast rotators $\lambda_R > 0.1$.
167: \citet{2007MNRAS.379..418C} showed that the slow and fast rotating galaxies of the SAURON sample populate distinct regions
168: in the $(v/\sigma,\epsilon)$ diagram, even if their inclination is taken into account, which means that those galaxies
169: are also intrinsically fast or slowly rotating.
170: Slow rotators tend to be more massive, have lower ellipticities and exhibit faster rotating, kpc-sized, old kinematically decoupled
171: components (KDCs), while fast rotators have lower luminosities and can also have central KDCs, which are however small
172: and often consists of young stellar populations \citep{2006MNRAS.373..906M}. Slow rotators also have strong kinematic twists
173: and kinematic misalignments, while photometry and velocity contours are almost always aligned in fast rotators.
174: In addition, a correlation between anisotropy and ellipticity has been detected for slow and fast rotators
175: with slow rotators being more isotropic and fast rotators being anisotropic (\citealp{2007MNRAS.379..418C}; \citealp{2007arXiv0710.0663B}).
176:
177: In this paper we want to connect the findings of the SAURON galaxies with their formation history. Ever since the work of
178: \citet{1972ApJ...178..623T} the merger of two late-type galaxies has been so far the most appealing and the best tested
179: theoretical model for the formation of an elliptical galaxy (\citealp{1992ApJ...393..484B}; \citealp{1996ApJ...471..115B};
180: \citealp{2003ApJ...597..893N}; \citealp{2006ApJ...650..791C}; \citealp{2006ApJ...641...21R}; \citealp{2006MNRAS.372..839N} and
181: references therein).
182:
183: The full two-dimensional kinematic field of merger remnants have scarcely been studied.
184: Following the pioneering work of \citet{2000MNRAS.316..315B}, \citet{2007MNRAS.376..997J}, henceforth J07,
185: analyzed kinematic maps of N-body merger remnants with kinemetric methods, introduced by \citet{2006MNRAS.366..787K}.
186: They quantified the presence of counter-rotating cores, the flattening of isovelocity contours,
187: counter-rotating discs and other peculiar features which are also observed in the SAURON galaxies. In
188: general, maps of slow rotators resembled more equal-mass mergers while
189: fast rotators show kinematical fields similar to unequal-mass mergers. However, with the
190: introduction of $\lambda_R$, we can quantify the line-of-sight angular momentum content of N-body
191: merger remnants in the same way as has been done with observed galaxies. Ideally we can connect merger
192: remnants which appear as slow rotator or fast rotator with a certain formation
193: mechanism. Beyond that we know the true angular momentum content of a simulated N-body merger remnant and can
194: therefore determine whether the remnant is indeed intrinsically a slow (fast) rotator or not.
195:
196: The aim of this work is two-fold: we want to study how much can be learned from the intrinsic structure of a
197: galaxy by calculating its line-of-sight angular momentum parameter $\lambda_R$. In addition, we want to
198: clarify whether remnants of binary
199: mergers follow the slow-fast rotator division of the red sequence as found in the SAURON sample.
200: The paper is organized as follows: we describe the simulations in Sec.\ref{sec:sims}, give details about
201: how we calculate the maps and determine $\lambda_R$ (Sec.\ref{sec:maps}) and examine
202: the connection between line-of-sight angular momentum and intrinsic angular momentum (Sec.\ref{sec:angmom}), show
203: the projectional behaviour of $\lambda_R$ for merger remnants with varying orbital content (Sec.\ref{sec:shape}),
204: classify the maps derived from the merger remnants as fast or slow rotators and investigate the influence of
205: various formation mechanisms (Sec.\ref{sec:slowfast}) and finally discuss the results in Sec.\ref{sec:discus}.
206: \section{Simulations}
207: \label{sec:sims}
208: \subsection{The Merger Models}
209: \label{MODELS}
210: The model parameters of the sample of 1:1 and 3:1 disc mergers we are studying
211: in this paper is identical to the sample presented in NJB2006, except
212: that we include star formation and stellar feedback in the modeling of the
213: dissipative component. Here we give a brief review of the setup.
214:
215: The disc galaxies were constructed in dynamical equilibrium using the
216: method described by \citet{1993ApJS...86..389H} with the following system of
217: units: gravitational constant G=1, exponential scale length of the
218: larger progenitor disc $h_d=1$ (the scale height was $h_z=0.2$) and
219: mass of the larger disc $M_d=1$. The discs were exponential with an
220: additional spherical, non-rotating bulge with mass $M_b = 1/3$, a
221: Hernquist density profile \citep{1990ApJ...356..359H} and a scale
222: length $r_b=0.2h_d$, and a pseudo-isothermal halo with a mass $M_d=5.8$,
223: cut-off radius $r_c=10h_d$ and core radius $\gamma=1h$. The parameters
224: for the individual components were the same as for the collisionless
225: mergers presented in NB03 and NJB2006. We replaced 10\% of the stellar
226: disc by gas with the same scale length and an initial scale height of
227: $h_{z,gas} = 0.1 h_z$.
228:
229: Whereas the NB2003 and NJB2006 simulations had been performed with
230: the Nbody-SPH code VINE \citep{2008arXiv0802.4253N,2008arXiv0802.4245W}
231: we used the parallel TreeSPH-code GADGET-2 \citep{2005MNRAS.364.1105S}
232: to follow radiative cooling \citep{1996ApJS..105...19K} for a primordial
233: mixture of hydrogen and helium together with a spatially uniform
234: time-independent local UV background \citep{1996ApJ...461...20H}. In addition,
235: we include star formation and the associated supernova feedback following
236: the sub-resolution multiphase model developed by \citet{2003MNRAS.339..289S}
237: using the same implementation as in \citet{2008arXiv0802.0210J} but without feedback from
238: black holes.
239:
240: We followed mergers of discs with mass ratios of 1:1 and 3:1.
241: The equal-mass mergers were calculated using in total 440000 particles
242: with each galaxy consisting of 20000 bulge particles, 60000 stellar
243: disc particles, 20000 SPH particles representing the gas component in the disc, and 120000 halo
244: particles. Twice as many halo particles than disc particles were used
245: to reduce heating and instability effects in the disc
246: components \citep{1999ApJ...523L.133N} by encounters between halo
247: and disc particles. For 3:1 mergers the parameters of the more
248: massive galaxy were as described above. The low-mass companion contained a third the mass and
249: the number of particles in each component, with a disc scale-length
250: (stars and gas) of $h=\sqrt{1/3 }$.
251:
252: The gravitational forces were softened with a Spline kernel of $h_{\mathrm{grav}} = 0.05$.
253: The minimal size of the Spline kernel used for computing the SPH
254: properties, $h_{\mathrm{SPH}}$, was fixed to the same
255: value. Implicitly, this procedure suppressed gas collapse on scales
256: smaller than the softening scale and prevents numerical instabilities
257: \citep{1997MNRAS.288.1060B}.
258: The initial discs were run in isolation for two dynamical times to allow
259: the systems to finally settle into an equilibrium state. The galaxies merged
260: on parabolic orbits with a pericenter distance of 2 disc scale lengths.
261: The initial geometries are identical to NB2003, Table 1. and NJB2006. The merger remnants were
262: allowed to settle into dynamical equilibrium for approximately 30 dynamical
263: time-scales after the merger was complete. Then their equilibrium state was analyzed.
264: All simulation have been run on a SGI Altix
265: at the University Observatory in Munich.
266:
267: \subsection{Determination of $\lambda_R$}
268: \label{sec:maps}
269: The two-dimensional maps examined in this article, are identically determined
270: as laid out in paper J07. We repeat here briefly the most important steps:
271: For the 2D analysis we binned particles within the central 6 length units (21 kpc) on a
272: grid of $48 \times 48$ cells. This corresponds typically to 2-3 effective
273: radii depending on projection (see \citealp{2006MNRAS.tmp..463N} for the exact
274: determination of $r_e$). To include seeing effects we created
275: for every luminous particle $10\times 10$ pseudo-particles with
276: identical velocities on a regular grid with a total size
277: of 0.125 unit lengths (0.44 kpc) centered on the original particle
278: position. The mass of the original particle was then distributed to
279: the pseudo-particles weighted by a Gaussian with a standard deviation
280: of 0.1625 unit lengths (0.56 kpc). Thereafter the pseudo-particles were binned
281: on a $48\times48$ grid.
282: For the kinematic analysis we binned (mass weighted) all pseudo-particles
283: falling within each grid cell in velocity along the line-of-sight. The width
284: of the velocity bins was set to a value of 0.1 (26.2 km/s) for line-of-sight velocities
285: $v_{\mathrm{los}}$ in the range $ -4 \le v_{\mathrm{los}} \le 4 $ (corresponding to $\pm 1048$ km/s).
286: This resulted in 80 velocity bins over the whole velocity interval. Using
287: the binned velocity data we constructed line-of-sight velocity
288: profiles (LOSVD) for each bin of the 2D grid. Subsequently we parameterized deviations
289: from the Gaussian shape of the velocity profile using Gauss-Hermite basis functions
290: \citep{1993MNRAS.265..213G,1993ApJ...407..525V}. The kinematic parameters of
291: each profile ($\sigma_{\mathrm{fit}}$, $v_{\mathrm{fit}}$, $h_3$, $h_4$)
292: were then determined by least squares fitting. \\
293: $\lambda_R$ can be calculated from integrated field data via the formula given
294: in EM07
295: \begin{equation}
296: \lambda_R=\frac{\Sigma_{i=1}^{N_p} F_i R_i |V_i|}{\Sigma_{i=1}^{N_p}F_i R_i \sqrt{V_i^2 +\sigma_i^2}},
297: \end{equation}
298:
299: where $F_i$ is the flux,$R_i$ the projected radius, $V_i$ the line-of-sight velocity and
300: $\sigma_i$ the line-of-sight velocity dispersion of each grid cell. We determined these properties as explained
301: in the previous section and can therefore calculate $\lambda_R$ directly from our 2D data. As galaxies are
302: projected randomly on the sky, we project each remnant 200 times randomly on the sphere of viewing directions. This
303: gives us a sample of 20000 2-dimensional maps. SAURON has a finite field-of-view which extends to one
304: effective radius, which we need to take into account as $\lambda_r$ can vary strongly with R. We determine for
305: each projection the effective radius and sum only over the grid cells inside one effective radius. Thus we ensure
306: a fair comparison to the SAURON data.
307: \section{The Intrinsic Specific Angular Momentum and $\lambda_R$}
308: \label{sec:angmom}
309: \subsection{Intrinsic Angular Momentum}
310: The total intrinsic specific angular momentum $\mathrm{J_{intr}}$ of a galaxy (or N-body merger remnant)
311: can be calculated from the discrete stellar (particle) distribution. The components of $\mathbf{J}$ in Cartesian
312: coordinates are
313: \begin{equation}
314: \mathrm{J}_x= \sum_i y_i v_{z,i} - z_i v_{y,i}
315: \label{eq:vec1}
316: \end{equation}
317: \begin{equation}
318: \mathrm{J}_y= \sum_i z_i v_{x,i} - x_i v_{z,i}
319: \label{eq:vec2}
320: \end{equation}
321: \begin{equation}
322: \mathrm{J}_z=\sum_i x_i v_{y,i} - y_i v_{x,i}
323: \label{eq:vec3}
324: \end{equation}
325: and $\mathrm{J}_{\mathrm{intr}}$ is the norm of that vector. For simplicity we implied here that all particles (stars) have
326: the same mass and do not write the sums over masses explicitly. The angular momentum vector can projected to the plane of
327: the sky, defining the {\it projected} intrinsic angular momentum vector $\mathbf{J_P}$ with components
328: \begin{equation}
329: \mathrm{J}_x= \sum_i y_i v_{z,i} - z_i v_{y,i}
330: \end{equation}
331: \begin{equation}
332: \mathrm{J}_y=0
333: \end{equation}
334: \begin{equation}
335: \mathrm{J}_z=\sum_i x_i v_{y,i} - y_i v_{x,i}
336: \end{equation}
337: where we define the y-direction arbitrarily as perpendicular to the plane of the sky. As our line-of-sight
338: is along the y-direction we can not observe transversal motions, i.e. in x and z-direction. We can now define
339: an apparent angular momentum vector $\mathbf{J_{app}}$ with the components
340: \begin{equation}
341: \mathrm{J}_x= \sum_i -z_i v_{y,i},
342: \end{equation}
343: \begin{equation}
344: \mathrm{J}_y=0,
345: \end{equation}
346: \begin{equation}
347: \mathrm{J}_z=\sum_i x_i v_{y,i}.
348: \end{equation}
349: It has been show by \citet{1988skeg.book.....F} that for triaxial galaxies the apparent
350: angular momentum amounts to half the intrinsic angular momentum, which we rewrite as
351: \begin{equation}
352: \label{eq:franx}
353: \mathbf{J_{P}}=\mathrm{\kappa_J} \mathbf{J_{app}},
354: \end{equation}
355: where $\kappa_J$ is typically 2. An additional complication is that the angular momentum of a
356: galaxy can be due both to streaming motions and to figure rotation. Again we follow \citet{1988skeg.book.....F}
357: \begin{equation}
358: \mathbf{J_{app}} = 0.5 \mathbf{J_{p}} + \mathrm{I_{proj}} \omega_\mathrm{proj},
359: \end{equation}
360: where $\mathrm{I_{proj}}$ and $\mathrm{\omega_{proj}}$ are the inertial moment tensor and the angular velocity
361: of the projected density distribution, respectively.
362: Indeed figure rotation is present in about 15\% of the merger remnants, but we do not try here to disentangle
363: the relative contributions of figure rotation and streaming motion. Although such a decomposition would be
364: possible in principle for the N-body remnants such an analysis would be beyond the scope of this paper.
365: \subsection{Angular Momentum in Observations}
366: We cannot directly access the apparent angular momentum of a galaxy observationally. The closest proxy to
367: the apparent angular momentum which we can calculate from the two-dimensional velocity field
368: is $\mathrm{<R|V|>}$, where R is the projected radius, V the line-of-sight-velocity and
369: brackets denote luminosity averages (see also Appendix A of EM07 for a discussion). It is
370: straightforward to calculate $\mathrm{J_{app}}$ from the N-body simulation and $\mathrm{<R|V|>}$ from mock
371: observations of the same remnants and same aperture. In Fig.\ref{fig:jproj} we show that agreement
372: is very good and verifies that the mapping of the velocity fields gives a close account of the
373: intrinsic structure.
374: \begin{figure}
375: \begin{center}
376: \epsfig{file=Japp_lambdaR_PAPER.eps,width=0.4 \textwidth}
377: \end{center}
378: \caption{The apparent angular momentum calculated from particle data (y-axis) as compared to the $\mathrm{<R|V|>}$ calculated
379: from the velocity maps (x-axis). The agreement is very good for all projections and all merger remnants. \label{fig:jproj}}
380: \end{figure}
381:
382: The inclination of a galaxy is in general not known and dynamical modeling or other observational
383: clues are needed to infer its projection \citep{2008arXiv0806.0042C}. We can choose the projection of an N-body remnant at will
384: and can calculate $\mathrm{J_{app}}$ and $\mathrm{J_{P}}$ for each projection. The ratio of these two values should be
385: equal to two, if and only if we have the information over the whole field of the galaxy. This is in general not the case, as the
386: field-of-view is limited by the instrument we are using. We are calculating therefore $\mathrm{J_{app}}$ and
387: $\mathrm{J_{P}}$ for two different apertures, one which encloses the whole remnant and one which extends only to
388: one effective radius. This is illustrated in the right plot of Fig.\ref{fig:kappaj} where we highlighted the aperture
389: at one effective radius.
390:
391: It is instructive to examine first a test model, for which we chose the progenitor disc galaxy, which
392: is by construction axisymmetric and fast rotating. For almost all projections we find that $\kappa_\mathrm{J}$
393: (Fig.\ref{fig:kappaj}, left) is very close to the theoretical value of 2, if calculated over the whole remnant. If we determine
394: $\mathrm{J_P}$ and $\mathrm{J_{app}}$ inside one $R_e$ we tend to get a spread of higher values for $\kappa_\mathrm{J}$. This is not too
395: surprising, as we catch the particles only on part of their trajectory (close to pericenter) and line-of-sight velocities are not
396: representative of their total angular momentum content. This will depend on the detailed orbit structure of the remnant, i.e.
397: how much angular momentum sits at large radii. In the same Fig.\ref{fig:kappaj} we illustrate the spread for the
398: whole remnant sample, which is somewhat larger than for our axisymmetric test models, but with the same tendency
399: to underestimate $\mathrm{J_P}$. Again using the information on the full remnants shows that the apparent angular momentum and the
400: projected intrinsic angular momentum are on average connected by a factor of two.
401: \begin{figure*}
402: \begin{center}
403: \epsfig{file=kappaj_PAPER.eps,width=0.6 \textwidth }
404: \epsfig{file=ISOPHOTE_PAPER.eps,width=0.3 \textwidth}
405: \end{center}
406: \caption{Left: Distribution of the ratio $\kappa_J=\mathrm{J_{P}}/\mathrm{J_{app}}$ for the axisymmetric
407: progenitor for 200 projections. The ratio shows the theoretically expected value of 2, if we calculate the angular
408: momentum terms for the whole remnant, while a larger spread of $\kappa_J$ is observed if we constrain the field-of-view
409: to one $R_e$. Middle: The same plot but this time for all collisionless remnants. Right: illustration of the size of the
410: aperture at one effective radius (white rhombus). \label{fig:kappaj}}
411: \end{figure*}
412:
413: The quantity $\lambda_R$, as defined by EM07, has been proposed as a proxy for the intrinsic
414: specific angular momentum of a galaxy within the radial range over which $\lambda_R$ is determined.
415: In the last test we want to see how closely $\lambda_R$ and the angular momentum content are
416: connected for a random projection. We discussed before that $\mathrm{<R|V|>}$, as a proxy
417: for $\mathrm{J_{app}}$, is closely connected to the projected intrinsic angular momentum. $\lambda_R$ is in
418: contrast to $\mathrm{<R|V|>}$ a normalized quantity. Therefore we scale both quantities to their
419: maximal values in order to compare the effect of inclination. We use again the progenitor disc galaxy as a test model
420: for which ellipticity maps inclination accurately. We see in Fig.\ref{fig:jintr} that $\lambda_R$ decreases much
421: less rapidly with inclination than $\mathrm{<R|V|>}$ does, which is also shown by the histograms of the deviations from their
422: maximal values (in this case scaled to one). The reason is that $<v>$ and $<\sigma>$ decrease
423: simultaneously and their ratio stays roughly constant until we reach inclinations close to face-on. The true angular momentum
424: content is of course most closely related to the maximum value of $\lambda_R$ or $\mathrm{<R|V|>}$, which in many cases, at least for
425: fast rotators is a projection close to edge-on. In summary, even if there is a close connection between $\mathrm{<R|V|>}$
426: and the projected intrinsic angular momentum $\mathrm{J_P}$, the projectional behaviour of $\lambda_R$
427: is more robust and will stay closer to its maximal value for a higher fraction of all possible
428: viewing angles.
429: \begin{figure*}
430: \begin{center}
431: \epsfig{file=PROJECT_DELTAJ_PAPER.eps,width=0.7 \textwidth}
432: \caption{Left: $\lambda_R$, $\mathrm{<R|V|>}$ and $\mathrm{J_{app}}$ for random projections of the progenitor galaxy.
433: All quantities are normalized to their maximum value. Right: Distribution of absolute deviation of the same quantities
434: from their (normalized) maximum values. \label{fig:jintr}}
435: \end{center}
436: \end{figure*}
437:
438: \section{Intrinsic Shape Variations: Example Remnants}
439: \label{sec:shape}
440: Before we examine the properties of the merger remnant sample as a whole it is instructive
441: to assess the variety of remnant properties present in the sample.
442: The remnants in our sample have very different orbital makeups
443: and intrinsic shapes. They can be dominated by major axis tubes, minor axis tubes
444: and box orbits, and have prolate, oblate or triaxial shapes respectively \citep{2005MNRAS.360.1185J}.
445: Their projected properties must correlate with their intrinsic shape and it is interesting to study
446: how $\lambda_R$ is varying with inclination. In Thomas et al. 2007 we selected
447: merger remnants at the extreme ends of the shape distribution and we want to
448: study the projected 2-dimensional properties of three of them to elucidate systematic kinematical trends
449: with intrinsic shape.
450:
451: Our test remnants are the OBLATE, PROLATE and TRIAXIAL remnants from \citet{2007MNRAS.381.1672T}.
452: We are plotting the ellipticity, $\lambda_R$ and the isophotal shape-parameter $a_4$ in Figure \ref{fig:examples}
453: to illustrate the variation with projection. The OBLATE remnant (red symbols) is the result
454: of a 4:1 merger,
455: which is dominated by minor axis tubes. It is classified as a fast rotator from almost any
456: viewing angle, although a dependence on inclination is visible. The TRIAXIAL remnant is a 1:1 merger remnant which is
457: dominated by box and boxlet orbits. Interestingly the remnant has a $\lambda_R$ consistent with 0
458: from all viewing angles. The PROLATE remnant (black symbols) has the most complex
459: dependency on inclination angle. This is so because it contains a significant amount of both minor and major axis tube
460: orbits. Therefore we almost always see some rotation, originating from different kinds of orbit classes (for the detailed description
461: of the orbit classification in N-body remnants see \citet{2005MNRAS.360.1185J}. However, as this
462: is also a 1:1 merger remnant it does not rotate very fast and reaches a maximum $\lambda_R$ of 0.18.
463:
464: The statistical distribution of $\lambda_R$ and $\epsilon$ are better seen in the right column of Fig \ref{fig:examples}.
465: The $\lambda_r$ of the TRIAXIAL remnant has a delta-function like distribution at $\lambda_R=0$, that means it is a
466: slow rotator under {\it any} viewing angle, while the OBLATE remnant is a fast rotator for almost all
467: projections except for three inclinations close to face-on. This results in a probability of $\approx 1 \%$ to
468: misclassify the OBLATE remnant as a slow rotator. The PROLATE remnant is one of the few remnants
469: which could be classified as a fast or a slow rotator with equal likelihood, because both tube orbit classes seem to carry
470: an equal amount of rotation and move along the line-of-sight for different viewing angles.
471: The PROLATE remnant is rather round for almost all projections, while the TRIAXIAL and the OBLATE remnant are quite
472: flattened, albeit the flattening is caused by different orbit classes, i.e. box orbits in the case of the TRIAXIAL
473: remnant. Flattened slow rotators are observed in the SAURON sample only in exceptional cases (see also Sec.\ref{sec:slowfast})
474:
475: \begin{figure*}
476: \epsfig{file=case_study_paper.eps,width=0.7 \textwidth}
477: \caption{Left: Case study of projectional relation between $\lambda_R$ and photometric properties, i.e. ellipticity and isophotal
478: shape parameter for an oblate (red), a prolate (black) and a triaxial remnant (blue). The triaxial remnant has
479: a $\lambda_R \approx 0$ not varying with projection. Ellipticity and $\lambda_R$ are good proxies for inclination angle for the oblate
480: remnant, which is nearly axisymmetric. The prolate remnant can have more than one $\lambda_R$ value for a given ellipticity
481: owing to a complicated orbital structure. Right: Histogram of ellipticities and $\lambda_R$ of the projectional study. The oblate remnant
482: is a fast rotator for almost all viewing angles, while the triaxial remnant is a slow rotator for all viewing angles. The prolate
483: remnant is equally probable to be observed as a fast or slow rotator. \label{fig:examples}}
484: \end{figure*}
485:
486:
487: \subsection{Kinematical Misalignment}
488: Another discriminatory property of slow and fast rotators is the kinematical misalignment. The misalignment angle
489: $\Psi$ is defined as the angle difference between the photometric and the kinematic position angle
490: \begin{equation}
491: \Psi = |\mathrm{PA}_\mathrm{phot} - \mathrm{PA}_\mathrm{kin}|,
492: \end{equation}
493: where $\mathrm{PA}_{\mathrm{kin}}$ is defined as the angle $0^{\circ}< \mathrm{PA}_{\mathrm{kin}} < 180^{\circ}$ along
494: which $|V|$ is maximal (see Appendix C of \citet{2006MNRAS.366..787K}). Triaxial systems can have complicated
495: velocity fields, because of the possible superposition of orbits with angular momenta with respect to the
496: minor and the major axis \citep{2008MNRAS.385..647V}.
497: In a perfect axisymmetric galaxy there can be no kinematic misalignment. Indeed all fast rotators in the SAURON
498: sample have misalignments $\Psi < 5^{\circ}$, whereas slow rotators can have any kinematic misalignment up to $90^{\circ}$.
499: Our example remnants are also quite diverse in their behaviour of kinematic misalignment with projection angle (Fig.\ref{fig:KM}).
500: The oblate remnant as it is dominated by minor axis tubes has very few projections where $\Psi \neq 0$. The triaxial
501: remnant has a much broader distribution, which peaks at low misalignments, but the maps of this remnant show little
502: rotation, such that the misalignment has larger error bars and is not so meaningful. The prolate remnant, however, shows
503: clearly a strong misalignment, i.e. $\Psi > 20$, for most of the projections, as both tube orbit classes revolving around
504: the major as well as the minor axis, are present in this remnant.
505: \begin{figure*}
506: \epsfig{file=case_study_KM_paper.eps,width=0.8 \textwidth}
507: \caption{Left: Case study of kinematical misalignment $\Psi$ of the same remnants as in Fig.\ref{fig:examples}.
508: Relation between $\lambda_R$ and kinematic misalignment. Right: Distribution of kinematical misalignments from 200
509: random projections. The OBLATE remnant has very few projections with $\Psi > 5^\circ$, while the PROLATE remnant distribution
510: of misalignment angles peaks at $40^\circ$. The TRIAXIAL remnant has a less broad distribution of $\Psi$ values.
511: \label{fig:KM}}
512: \end{figure*}
513:
514: \section{Slow and Fast Rotators}
515: \label{sec:slowfast}
516: The determination of the $\lambda$-parameter is important in sofar as ellipticals can be divided
517: into two broad sub-classes, so called slow and fast rotators (EM07). The (empirical) cut which separates
518: these classes has been made at $\lambda_R=0.1$. These seem to be genuinely intrinsically different galaxy
519: types with little overlap through projection effects. Many slow rotators exhibit e.g. large, old and
520: fast rotating kinematically decoupled components (KDCs), which however are not affecting too much the
521: overall angular momentum balance, as the stars which move on high angular momentum orbits are not located in
522: the center, such that the $\lambda_R$ as determined from the map is rather low. In general maps of slow
523: rotators resemble more the kinematic maps derived from equal-mass merger remnants. The formation
524: of KDCs requires the inclusion of a gaseous component in the simulation. Fast rotators are
525: on the other hand more likely to be un-equal mass mergers. For a detailed discussion on the origin
526: of kinematical features in 2-dimensional LOSVD maps we refer the reader to J07.
527:
528: \subsection{The Role of Merger Mass Ratios}
529: \label{sec:massratio}
530: The most fundamental parameter which influences the outcome of a merging event is the amplitude of the fluctuations
531: of the gravitational potential during the merger. How strongly the phase space is rearranged after a merger
532: will be mainly determined by the ratio of the masses of the merging objects \citep{2003ApJ...597..893N}. An equal-mass
533: merger will be more effective in redistributing the energies of the stars in the progenitor galaxies than an un-equal mass merger.
534: It is true that secondary factors like infall velocity and merging geometry have also an impact on the final
535: shape of the remnant, but the mass ratio of the merger is the dominant factor.
536: We want to quantify when a given projection of a
537: given remnant is classified as a slow or a fast rotator and we use collisionless disc-disc merger remnants
538: of mergers with mass ratios of 1:1, 2:1, 3:1 and 4:1 from \citet{2003ApJ...597..893N}.
539: In Fig.\ref{fig:massratio} the $\lambda_R$ distribution is shown for each of these merger subsamples. We see, as expected,
540: that the fraction of slow rotators for a given merger mass ratio is decreasing with increasing mass ratio.
541: \begin{figure}
542: \epsfig{file=lambda_eta_PAPER.eps,width=0.45 \textwidth}
543: \caption{Distribution of $\lambda_R$ for collisionless merger remnants of different merger mass ratios. The region
544: where kinematic maps are classified as slow rotators is shaded. Only binary mergers of equal-mass progenitors
545: seem to produce slow rotators in sizeable fractions. \label{fig:massratio} }
546: \end{figure}
547: $\lambda_R$ peaks at 0, 0.1, 0.27 and 0.4 for the mass ratios 1:1 to 4:1. If one would assume that binary
548: disc-merging is the main formation channel of elliptical galaxies then only equal or near equal-mass merger could produce
549: them. However, this picture is probably too simple as elliptical galaxies can have more complicated merger histories in a cosmological
550: context (\citealp{2003ApJ...590..619M};\citealp{2007ApJ...658..710N}). Multiple minor mergers, e.g. three 3:1 mergers might
551: also be a viable way to form slowly rotating ellipticals, which resemble closely the
552: outcome of a single equal-mass merger, as was shown recently by \citet{2007A&A...476.1179B}. All the results (except
553: for Sec.\ref{sec:dry}) we present in this paper are valid for one generation of merging only. The examination of the impact of
554: complex hierarchical merging histories is beyond the scope of this paper.
555: \begin{table}
556: \begin{center}
557: \begin{tabular}{cc}
558: \hline\hline
559: Merger Mass Ratio & Slow Rotator Fraction \\
560: \hline
561: 1:1 & 0.75 \\
562: 2:1 & 0.10 \\
563: 3:1 & 0.03 \\
564: 4:1 & 0.02 \\
565: \hline
566: \end{tabular}
567: \caption{Dependence of the slow rotator fraction on the merger mass ratio. Only equal-mass merger are very effective
568: in removing angular momentum in a single merging event. \label{tab:massratio} }
569: \end{center}
570: \end{table}
571: \subsection{Confusion Rate}
572: \citet{2007MNRAS.379..418C} showed that slow and fast rotators are indeed intrinsic fast and slow rotators
573: and not randomly projected fast rotating galaxies which happen to appear non-rotating. We address this
574: question in Fig. \ref{fig:confuse} where we examine the fraction of projections for each of the collisionless
575: remnants which have $\lambda_R < 0.1$. There is an isolated sample of remnants which have a probability of greater
576: than 90\% to be classified as slow rotators. The remaining remnants have a probability of less than 40\% to
577: be identified as slow rotators. The mergers which are located at the extreme ends of the distribution are the
578: unambiguous cases they are (almost) always either fast rotators or slow rotators.
579: There is however a small population of remnants which have a reasonable probability to be identified
580: as slow rotators although they have a non-zero angular momentum content (shaded zone in
581: Fig.\ref{fig:confuse}, left). These are on average the most prolate of all remnants
582: (same Figure, right plot). The confusion arises from the presence of major and minor axis tubes
583: in the remnants, i.e. there are two face-on projections
584: under which the angular momentum content is perpendicular to the line-of-sight and cannot be observed.
585: But there also remnants which are mildly to strongly triaxial ($0.3 < T < 0.5$), which are hot but regularly
586: rotating systems, i.e. with low kinematic misalignment. Almost all of these systems resulted from
587: 2:1 mergers. We can calculate now the global probability to falsely classify a rotating galaxy
588: as a slow rotator in our sample and find $\mathrm{P_{confuse}}=0.046$. This is probably an upper limit as
589: prolate remnants are overabundant in our sample, and are suppressed if a significant amount of
590: gas is included in the simulations \citep{2006MNRAS.372..839N}. This is consistent with the SAURON sample
591: where not a single galaxy has been observed with rotational patterns as, e.g. our PROLATE model, which
592: we presented in Sec.\ref{sec:shape}. We further note that the maximal $\lambda_R$ is a good tracer
593: for the triaxiality of our remnants (Fig.\ref{fig:confuse},right), albeit with an increasing spread
594: towards lower $\lambda_R$. Particular spin alignments of the progenitor discs can lead to interesting
595: outliers from this correlation, e.g. we find a slow rotator with $T \approx 0$. This merger remnant
596: originated from a planar merger of discs with anti-aligned spins. The counter-rotation of both disc
597: components cause a net zero rotation. While this is a possible formation scenario, most slow rotators
598: are triaxial box dominated remnants.
599:
600: \begin{figure*}
601: \epsfig{file=CONFUSE_PAPER.eps,width=0.8 \textwidth}
602: \caption{Left: Distribution of mergers according to their probability of being classified
603: as a slow rotator ($\lambda_R < 0.1$). Right: Correlation between the maximal $\lambda_R$ and the
604: intrinsic triaxiality T of the merger remnants. Open symbols indicate merger remnants which belong
605: to the shaded area from the plot on the left side, i.e. have a significant number of misclassified
606: projections. \label{fig:confuse}}
607: \end{figure*}
608:
609: \subsection{The Influence of Dissipation}
610: Late-type spiral galaxies are normally not purely collisionless systems but have a sizable fraction of gas.
611: We therefore compare two sets of mergers: disc-disc mergers with gas and star formation, and collisionless
612: disc-disc mergers. Each set consists of 16 1:1 and 32 3:1 mergers, which formed on identical merging geometries. As
613: mentioned before we take 200 observations of each remnant at random viewing angles, which results in 9600 maps for each
614: set of simulations.
615: We determine for each set the fraction of slow and fast rotators, as well as standard photometric parameters such as
616: ellipticity and the isophotal shape parameter $a_4$. In Fig. \ref{fig-1} we show the distributions of ellipticities
617: and $a_4$ for slow and fast rotators from mergers which formed with and without gas. Maps classified as slow and
618: fast rotators show distinct photometric properties. Slow rotators in general have smaller ellipticities
619: and have smaller $a_4$ values while fast rotators are more elliptical and discy.
620: It is apparent that the fast rotator population hardly changes its photometric properties when gas is
621: included in the simulation.
622: The impact on slow rotators, however, is more visible as the ellipticity distribution is now more skewed towards $\epsilon=0$,
623: while the $a_4$ distribution peaks now at 0 (\citealp{2006ApJ...650..791C};\citealp{2006MNRAS.372..839N}).
624: The mean slow rotator properties are summarized in Table \ref{tab-1}, which also confirm that most
625: slow rotators originate from equal-mass mergers.
626: \begin{figure}
627: \begin{center}
628: \epsfig{file=global_histo_PAPER.eps,width=0.5 \textwidth}
629: \caption{Distribution of ellipticities and isophotal shape parameter for the slow rotator (left plots)
630: and the fast rotator (right plots) population. The collisionless 1:1 and 3:1 remnants are indicated by
631: black solid lines. Fast and slow rotators have clearly separate photometric
632: properties. Fast rotators have high ellipticities and discy isophotes. Histograms for dissipatively formed
633: merger remnants are overplotted in (dashed) red. Fast rotators are very little affected while slow rotators
634: with gas have significantly lower ellipticities, i.e. are rounder. \label{fig-1} }
635: \end{center}
636: \end{figure}
637: The results from the SAURON survey showed that slow and fast-rotating ellipticals are located at certain positions in
638: the $\lambda -\epsilon$ and in the $\lambda - a_4$ plane. For a better comparison we
639: also show the distribution of merger remnants in those parameter spaces (Fig. \ref{fig-2}). In general the parameter space
640: bracketed by the SAURON observations is reproduced well by the merger remnants. There are, however, some differences.
641: For example
642: we see that the collisionless mergers have too high $\epsilon$ for a given $\lambda_R$ which is alleviated, especially
643: for slow rotators, by the inclusion of gas. We also indicated the progenitor galaxy in Fig.\ref{fig-2}. We already showed
644: in Sec.\ref{sec:massratio} that the amount of rotational support always decreases after a merger event.
645: Very high merger mass ratios will lead to merger remnants which would lie closer to the progenitor limit, but we have not performed
646: 10:1 or even higher mass ratios. Therefore we do not find projections of merger remnants with $\lambda_R > 0.6$. It is remarkable
647: though that some of the SAURON fast rotators have ($\lambda_R$,$\epsilon$)-values very close to the progenitor galaxy
648: and are, at least in this sense, hardly distinguishable from early-type spiral galaxies.
649:
650: The situation is different in the $a_4 - \lambda$-plane where the merger remnants are only partly
651: located in the parameter space occupied by the SAURON galaxies. Collisionless slow rotators are too boxy while fast rotators
652: are on average too discy compared to SAURON galaxies. While gas removes the boxy projections of slow rotators it does not help
653: to create, as expected, boxy projections in fast rotators. Probably the origin of boxiness in the
654: fast rotating SAURON galaxies is different from the source of boxiness in our remnants, e.g. bars.
655: \begin{figure*}
656: \epsfig{file=lambda_ell_2D_PAPER.eps,width=0.85 \textwidth }\\
657: \epsfig{file=lambda_a4_2D_PAPER.eps,width=0.85 \textwidth }
658: \caption{Left Column: Two-dimensional probability distribution of collisionless 1:1 and 3:1 remnants in $\lambda_R - a_4$,
659: respectively in the $\lambda_R - \epsilon$ plane. The SAURON galaxies are indicated by black stars. High
660: $\lambda_R$ galaxies are resembling more the progenitor galaxy than 3:1 merger remnants. Boxy fast rotators
661: are not formed in mergers, but are probably barred galaxies. Right Column:
662: The same plots but this time with merger remnants which formed with gas. The photometric properties
663: are stronger affected than $\lambda_R$. Slow rotating remnants agree well with
664: SAURON galaxies.\label{fig-2}}
665: \end{figure*}
666:
667: As we discussed earlier, kinematical misalignment is also a distinguishing feature between fast and slow rotators.
668: Again we divide up the simulation sample into fast and slow rotators and measure the angle $\Psi$ between kinematic
669: and photometric position angles. The observational trends are reproduced remarkably well: fast rotators
670: are very unlikely to show kinematic misalignment, while slow rotators show a flat distribution of misalignment
671: angles, in other words there is no preferred alignment angle for slow rotators (see Fig.\ref{fig:kmhisto}).
672: The low misalignment angles for fast rotators are not very surprising, because we already start out with
673: a perfectly aligned system. Unequal mass merger simply do not destroy the initial alignment. Indeed observed fast
674: rotators are also remarkably aligned. Only 2 out of 50 fast rotators have $\Psi > 5^\circ$ (see \citealp{2007MNRAS.379..418C}, Fig.12). The
675: crucial question to explain the origin of fast rotators is rather how the axisymmetric progenitor systems are
676: produced in the first place, rather than the subsequent merging.
677:
678: Gas changes the picture only slightly. There are somewhat fewer projections with high misalignment angles for fast
679: rotators and slow rotators are slightly better aligned. In general the inclusion of only 10\% of gas in the merging
680: process is not necessary to explain the kinematic misalignment of slow and fast rotators. However, the inclusion of high
681: fractions of gas would probably change this picture.
682: \begin{figure}
683: \epsfig{file=psi_histo_PAPER.eps,width=0.5 \textwidth }
684: \caption{Histograms of kinematic misalignment angle $\Psi$ measured from 2-dimensional kinematical maps
685: of the merger remnants. The overwhelming majority of maps identified as fast rotators have no
686: kinematic misalignment, while slow rotators have no preferred misalignment angle. Dissipation changes
687: the results only slightly, i.e. the maps slow rotators have appear slightly more aligned. \label{fig:kmhisto}}
688: \end{figure}
689:
690: \begin{table}
691: \begin{center}
692: \begin{tabular}{lcccccc}
693: \hline\hline
694: & Collisionless & Gas + SF \\
695: \hline
696: Slow Rotator Fraction & 0.26 & 0.28 \\
697: Slow Rotator $<\epsilon>_{med}$ & 0.28 & 0.18 \\
698: Slow Rotator $\epsilon < 0.3$ & 0.55 & 0.81 \\
699: Slow Rotators from 1:1 merger & 0.95 & 0.94 \\
700: \hline
701: \end{tabular}
702: \caption{Slow Rotator Properties. Main results of the classification according to the $\lambda_R$-parameter. Every
703: projection which has a $\lambda < 0.1$ is classified as a slow rotator. Dissipation mainly influences the ellipticity, while
704: the slow rotator fraction is almost unaffected. \label{tab-1} }
705: \end{center}
706: \end{table}
707:
708: \subsection{Dry Mergers}
709: \label{sec:dry}
710: The most massive galaxies in the universe are probably not formed from a single
711: generation of binary spiral-spiral mergers \citep{2007astro.ph..2535N}, because there are simply not enough massive late-type
712: galaxies to account for the stellar mass of galaxies with luminosities, e.g., of 4 $L^*$. Their merging history included
713: probably more than one major merger and semi-analytic modeling showed that the last major merger
714: was probably dry (\citealp{2003ApJ...597L.117K};\citealp{2007ApJ...659..976H}). We have a small sample
715: of re-mergers of collisionless merger remnants to test the assumption if slow rotators are the end product
716: of dry merging. We analyzed 6 equal-mass mergers of discy remnants and 6 equal-mass mergers of boxy remnants.
717: The majority of the maps (about 68 \%) of the dry mergers are indeed slow rotators. As we have a very limited
718: sample of merger remnants, we compare them only with the slow rotators which originated from 1:1 collisionless
719: disc mergers, which is admittedly the overwhelming majority. In Fig. \ref{fig-dry} we show that the ellipticity
720: distribution from elliptical-elliptical mergers is hardly distinguishable
721: from equal-mass disc-disc mergers. However, they are on average more boxy than remnants of
722: disc mergers in agreement with \citet{2006ApJ...636L..81N}. The slow rotators in the SAURON sample are indeed
723: hardly boxy. As mentioned before, the SAURON sample is not representative of the shape distribution of all elliptical
724: galaxies due to the selection procedure of galaxies. There is especially a dearth of triaxial, boxy galaxies
725: which makes it difficult to exclude or confirm the dry merging origin of massive ellipticals with the present
726: sample (see also \citet{2007arXiv0710.0663B} for a more detailed discussion of the origin of the slowly rotating
727: SAURON ellipticals).
728:
729: \begin{figure}
730: \epsfig{file=dry_histo_PAPER.eps,width=0.5 \textwidth}
731: \caption{Distribution of photometric properties for slow rotators from collisionless 1:1 elliptical and 1:1 disc mergers.
732: The ellipticity distribution of both samples are similar while elliptical merger remnants are more boxy. \label{fig-dry} }
733: \end{figure}
734:
735: \section{Discussion and Conclusions}
736: \label{sec:discus}
737: We calculated the $\lambda_R$-parameter from line-of-sight velocity maps of 48 remnants of collisionless disc
738: mergers, 48 remnants of disc mergers with a dissipative component and
739: 12 re-mergers of collisionless remnants, according to the procedure laid out in
740: EM07. Every remnant is projected 200 times randomly and then treated as a mock galaxy.
741: We first investigated how the intrinsic properties of the merger remnants manifest itself in the maps and how much
742: can be deduced of the intrinsic structure by calculating $\lambda_R$. We then identified a mock galaxy as a fast or
743: slow rotator and compare their photometric and kinematic properties like $a_4$, $\epsilon$ and kinematic misalignment
744: $\Psi$ to the properties of SAURON galaxies. We examined the influence of gas and merging mass ratio on the statistical
745: properties of slow and fast rotators.
746: We report the following findings:
747: \begin{itemize}
748: \item
749: The line-of-sight angular momentum parameter $\lambda_R$ is a rather robust indicator of the
750: intrinsic angular momentum content of a galaxy. $\lambda_R$ stays close to its maximum value for a wide range of
751: viewing angles as it captures the balance between ordered (line-of-sigh velocity) and unordered motion
752: (line-of-sight velocity dispersion) well. In principle the projected angular momentum content could also be deduced from the
753: two-dimensional velocity field alone, but information over the entire field of the galaxy would be needed, which is
754: normally not available. We conducted therefore tests to explore the deviation from the intrinsic angular
755: momentum, observing galaxies with apertures of limited size.
756: \item
757: The projectional behaviour of $\lambda_R$ is determined by the intrinsic orbital structure.
758: Most fast rotators consist of disc-like minor axis tubes and have $\lambda_R > 0.1$ for almost all projections.
759: Likewise slow rotators consist of box orbits and have $\lambda_R < 0.1$ for all projections. Prolate remnants have
760: a more complicated projectional behaviour as they consist of major and minor axis tubes. There is however not a single
761: galaxy in the SAURON survey with significant rotation around the major axis and we slightly overproduce these
762: systems in our merger sample. Even so we determine the misclassification probability for the whole collisionless
763: merger sample to only 4.6 \%. We therefore confirm the robustness of the slow/fast rotator classification scheme
764: for binary disc merger remnants.
765: \item
766: The spread of $\lambda_R$ is naturally reproduced in the binary merger scenario. However their photometric
767: properties have some important differences. Slow rotators from collisionless mergers have too high ellipticities. The
768: problem is alleviated by the inclusion of a modest fraction of gas, which does not modify the slow rotator fraction
769: but produces remnants with lower ellipticities. Therefore gas, maybe counter-intuitively, can be important for
770: the formation of slow rotators.
771: \item
772: The destruction of ordered rotation is only effective in equal-mass mergers. The violence of the merger seems
773: to be the decisive factor, i.e. with increasing mass-ratio the percentage of formed slow rotators drops
774: very rapidly. This is true at least for the binary merger scenario, very large gas fractions and multiple mergers will
775: complicate this picture.
776: \item
777: We tested the amount of kinematic misalignment for slow and fast rotators separately. Again the agreement
778: with observations is very good. Slow rotators can have any misalignment with equal probability, while fast rotators
779: show kinematical alignment for almost all projections. The picture changes with the inclusion of gas only insofar as the
780: strongest misalignments vanish for slow rotators. Higher gas fraction will probably lead to a further decrease in observed
781: misalignments.
782: \item
783: Slow rotators in the SAURON sample are not strongly boxy and rather round, therefore the dry mergers
784: are a poor fit to them. This is somewhat surprising, as the projection and mass ratio independence of
785: boxiness was seen as an asset of this formation mechanism \citep{2006ApJ...636L..81N}.
786: To assess the role of dry mergers adequately probably a larger sample of ellipticals, like ATLAS$^{\rm 3D}$, is needed.
787: \item
788: Fast rotators with boxy isophotes are absent from our sample. Some early-type S0s in the SAURON
789: are barred galaxies, and such strong bars seem not to be formed in our particular merger sample. Our sample
790: of course has its limits and it is possible that the bulge to disc ratio of our progenitor galaxies is too high
791: or the initial bulges are not rotating or probably both, to compare to this particular property of the SAURON galaxies.
792: \item
793: The fast rotators with the highest $\lambda_R$ values in the SAURON sample resemble our progenitor galaxy more
794: than a merger remnant, which leads to the question, if major merging played a significant role in their formation
795: at all.
796: \end{itemize}
797:
798: \section*{Acknowledgments}
799: This research was supported by the DFG priority program SPP 1177 and by the DFG
800: cluster of excellence 'Origin and Structure of the Universe'. Part of the simulations
801: were run on the local SGI ALTIX 3700 Bx2 which was also partly funded by this cluster of excellence.
802: MC acknowledge support from a STFC Advanced Fellowship (PP/D005574/1).
803: \bibliographystyle{mn2e}
804: \bibliography{references}
805:
806:
807: \clearpage
808:
809:
810:
811:
812: \clearpage
813:
814: %% Use the figure environment and \plotone or \plottwo to include
815: %% figures and captions in your electronic submission.
816: %% To embed the sample graphics in
817: %% the file, uncomment the \plotone, \plottwo, and
818: %% \includegraphics commands
819: %%
820: %% If you need a layout that cannot be achieved with \plotone or
821: %% \plottwo, you can invoke the graphicx package directly with the
822: %% \includegraphics command or use \plotfiddle. For more information,
823: %% please see the tutorial on "Using Electronic Art with AASTeX" in the
824: %% documentation section at the AASTeX Web site,
825: %% http://www.journals.uchicago.edu/AAS/AASTeX.
826: %%
827: %% The examples below also include sample markup for submission of
828: %% supplemental electronic materials. As always, be sure to check
829: %% the instructions to authors for the journal you are submitting to
830: %% for specific submissions guidelines as they vary from
831: %% journal to journal.
832:
833: %% This example uses \plotone to include an EPS file scaled to
834: %% 80% of its natural size with \epsscale. Its caption
835: %% has been written to indicate that additional figure parts will be
836: %% available in the electronic journal.
837:
838:
839:
840:
841:
842:
843:
844: %% This figure uses \includegraphics to scale and rotate the still frame
845: %% for an mpeg animation.
846:
847: %\begin{figure}
848: %\includegraphics[angle=90,scale=.50]{f3.eps}
849: %\caption{}
850: %\end{figure}
851:
852:
853:
854:
855: \clearpage
856:
857:
858:
859: %% If the table is more than one page long, the width of the table can vary
860: %% from page to page when the default \tablewidth is used, as below. The
861: %% individual table widths for each page will be written to the log file; a
862: %% maximum tablewidth for the table can be computed from these values.
863: %% The \tablewidth argument can then be reset and the file reprocessed, so
864: %% that the table is of uniform width throughout. Try getting the widths
865: %% from the log file and changing the \tablewidth parameter to see how
866: %% adjusting this value affects table formatting.
867:
868: %% The \dataset{} macro has also been applied to a few of the objects to
869: %% show how many observations can be tagged in a table.
870:
871:
872:
873: \end{document}
874:
875: %%
876: %% End of file `sample.tex'.
877: