1: %\documentclass[useAMS,usenatbib]{mn2e}
2: %\documentclass[useAMS,usenatbib,usegraphicx,onecolumn]{mn2e}
3: \documentclass[useAMS,usenatbib,usegraphicx,twocolumn]{mn2e}
4: %\documentstyle[referee,epsfig]{mn}
5: %\documentstyle[onecolumn,epsfig]{mn}
6: \voffset -1cm
7:
8: \def\be{\begin{eqnarray}}
9: \def\ee{\end{eqnarray}}
10:
11: \def\h{_{\rm h}}
12: \def\H{_{\rm H}}
13: \def\bh{_{\rm BH}}
14: \def\sph{_{\rm sph}}
15: \def\nsph{_{\rm nsph}}
16: \def\calJ{{\cal J}}
17:
18: \def\dex{{\rm\,dex}}
19: \def\kms{{\rm\,km\,s^{-1}}}
20: \def\Lsun{{\rm\,L_\odot}}
21: \def\msun{{\rm\,M_\odot}}
22: \def\mag{{\rm\,mag}}
23: \def\Mpc{{\rm\,Mpc}}
24: \def\kpc{{\rm\,kpc}}
25: \def\yr{{\rm\,yr}}
26: \def\Myr{{\rm\,Myr}}
27: \def\pc{{\rm\,pc}}
28: \def\cm{{\rm\,cm}}
29: \def\au{{\rm\,AU}}
30:
31: \def\bulge{_{\rm bulge}}
32: \def\disk{_{\rm disk}}
33: \def\sub{{_{\rm sub}}}
34: \def\halo{_{\rm halo}}
35: \def\vir{_{\rm vir}}
36:
37: \title[Kinematics of hypervelocity stars]{Kinematics of hypervelocity stars in the triaxial halo of the Milky Way}
38: \author[Q.\ Yu \& P.\ Madau]{Qingjuan Yu\thanks{Also a Hubble Fellow at the Department of Astronomy, University of California at Berkeley, Berkeley, CA 94720.}
39: and Piero Madau\thanks{Also at the Max-Planck-Institut f\"ur Astrophysik,
40: Karl-Schwarzschild-Str. 1, 85740 Garching, Germany.}\\
41: Department of Astronomy and Astrophysics, University of California, Santa Cruz, CA 95064\\
42: yqj@ucolick.org; pmadau@ucolick.org}
43:
44: \begin{document}
45:
46: \label{firstpage}
47: \maketitle
48:
49: \begin{abstract}
50: Hypervelocity stars (HVSs) ejected by the massive black hole at the
51: Galactic center have unique kinematic properties compared to other halo stars. Their
52: trajectories will deviate from being exactly radial because of the asymmetry
53: of the Milky Way potential produced by the flattened disk and the triaxial dark
54: matter halo, causing a change of angular momentum that can be much larger than
55: the initial small value at injection. We study the kinematics of HVSs and
56: propose an estimator of dark halo triaxiality that is determined only by
57: instantaneous position and velocity vectors of HVSs at large Galactocentric
58: distances ($r\ga 50\,\kpc$). We show that, in the case of a substantially triaxial halo,
59: the distribution of deflection angles (the angle between the stellar
60: position and velocity vector) for HVSs on bound orbits is spread uniformly
61: over the range 10$^\circ$--180$^\circ$. Future astrometric and deep wide-field
62: surveys should measure the positions and velocities of a significant number of
63: HVSs, and provide useful constraints on the shape of the Galactic dark matter
64: halo.
65: \end{abstract}
66: \begin{keywords}
67: black hole physics -- Galaxy: center -- Galaxy: halo -- stellar dynamics
68: \end{keywords}
69:
70: \section{Introduction}\label{sec:intro}
71:
72: Recent observations have revealed the existence of a population of
73: hypervelocity stars (HVSs) traveling in the halo of the Milky Way (MW) with
74: Galactic rest-frame velocities $v_{\rm rf}$ in the range between $+400$ and $+750\,\kms$
75: \citep{Brown05,Edelmann05,Hir05,Brown06a,Brown06b}. HVSs are probably B-type
76: main sequence stars with lifetimes $\la 100\,\Myr$, Galactocentric distances
77: $>50$ kpc, and move with speeds large enough to escape from the Galaxy.
78: The significant excess of B-type stars with velocities $+275<v_{\rm
79: rf}<+450\,\kms$ and distances $>10$ kpc observed by \citet{Brown07} may also
80: be an indication that many HVSs are ejected into the halo on {\it bound} orbits.
81:
82: HVSs were first recognized by \citet{H88} as an unavoidable byproduct of the presence
83: a massive black hole (BH) in the Galactic center. Only a close encounter with
84: a relativistic potential well can accelerate a 3-4 $\msun$ star to such extreme
85: velocities, and at least three different ejection mechanisms have been
86: proposed: the interaction between background stars and an intermediate-mass black
87: hole (IMBH) inspiralling towards Sgr A$^*$ \citep{YT03,Levin06,BGPZ06,SHM06}, the
88: disruption of stellar binaries in the tidal field of Sgr A$^*$
89: \citep{H88,YT03,GL06,Bromley06}, and the scattering of stars off a cluster
90: of stellar-mass BHs orbiting Sgr A$^*$ \citep{OL06}
91: In all these models,
92: HVSs have unique kinematics compared to other halo stars: 1) they have almost zero
93: initial specific angular momentum at ejection, $\sqrt{GM\bh r_p}
94: \simeq 4.0\times 10^{-6}\kpc^2\Myr^{-1} (M\bh/3.6\times 10^6\msun)^{1/2}
95: (r_p/10^{-6}\kpc)^{1/2}$, where $M\bh$ the mass of Sgr $A^*$ and $r_p$
96: the pericenter distance of the star;
97: 2) their high speeds diminish the impact of two-body relaxation or dynamical
98: friction effects on their motion; and 3) their trajectories will deviate from
99: being exactly radial because of the asymmetry of the Milky Way potential
100: produced by the flattened disk and the triaxial dark matter (DM) halo, causing a
101: change of angular momentum that can be much larger than the initial small value.
102: (For reference, a $1\,\kms$ deviation of the velocity from the radial
103: direction at $50\kpc$ represents a change of $0.05\kpc^2\Myr^{-1}$ in specific
104: angular momentum.) Proper-motion measurements of HVSs may therefore become
105: a key diagnostic tool for constraining the shape of the Galactic
106: potential \citep{Gnedin05}.
107:
108: Triaxial halos are a generic prediction of the hierarchical, cold
109: dark matter (CDM) models of structure formation. Dissipationless cosmological
110: simulations typically predict minor-to-major density axis ratios in the range
111: 0.4-0.8 (e.g. \citealt{JS02}), with the asphericity of the potential increasing
112: rapidly towards the center of the halo \citep{HNS06}. Gas cooling tends
113: to circularize the potential (e.g. \citealt{Dub94,Kazant04}), while subsequent
114: mergers produce highly elongated remnants (e.g. \citealt{Moore04}).
115: Studies of weak gravitational lensing and X-ray observations of elliptical
116: galaxies show that halos are significantly flattened, in fair agreement with results
117: from numerical simulations \citep{Hoe04,Buo02}. Yet the coherence of tidal debris
118: from the Sagittarius dwarf galaxy appears to indicate that the inner halo of
119: the MW is nearly spherical and therefore in conflict with CDM predictions
120: (Ibata et al. 2001; but see Helmi 2004).
121:
122: In this paper, we study the kinematics of HVSs in the MW as a probe of the
123: triaxiality of the Galactic halo. The outline is as follows. In
124: \S~\ref{sec:analysis}, we analyze the motion of HVSs in a flattened or
125: triaxial gravitational potential. We provide a
126: concise statistical estimator for the triaxiality of the Galactic halo
127: potential through the measured angular momenta of HVSs. In \S~\ref{sec:potential}, we
128: review the Galactic potential model to be used in our calculations. In
129: \S~\ref{sec:simulation} we perform numerical simulations of the motion of HVSs to
130: study their kinematics. Finally, in \S~\ref{sec:conclusion}, we summarize our
131: conclusions.
132:
133: \section{Motion of hypervelocity stars}\label{sec:analysis}
134:
135: Consider a star with position vector $\vec{r}$ moving with velocity $\vec{v}$ in
136: a gravitational potential
137: $\Phi=\Phi\sph(r)+\Phi\nsph(x,y,z)$, where $\Phi\sph$ and $\Phi\nsph$ are
138: the spherically-symmetric and aspherical component of the the potential,
139: $(x,y,z)$ are Cartesian coordinates, and $r=\sqrt{x^2+y^2+z^2}$. The rate of
140: change of the specific angular momentum of the star, $\vec{J}=\vec{r}\times \vec{v}$,
141: is equal to the torque,
142: \be
143: d\vec{J}/dt=-\vec{r}\times\nabla\Phi=-\vec{r}\times\nabla\Phi\nsph,
144: \label{eq:dJdt}
145: \ee
146: and has components
147: \be
148: \cases{
149: dJ_x/dt=2yz\left(\partial\Phi\nsph/\partial y^2-\partial\Phi\nsph/\partial z^2\right), \cr
150: dJ_y/dt=2xz\left(\partial\Phi\nsph/\partial z^2-\partial\Phi\nsph/\partial x^2\right), \cr
151: dJ_z/dt=2xy\left(\partial\Phi\nsph/\partial x^2-\partial\Phi\nsph/\partial y^2\right).
152: }
153: \label{eq:dJidt}
154: \ee
155: It is convenient to change from Cartesian to spherical coordinates,
156: $(x,y,z)=(r\sin\theta\cos\phi,r\sin\theta\sin\phi,r\cos\theta)$,
157: and combine the above equations to yield
158: \be
159: \frac{dJ_x/dt}{\sin\theta\sin\phi\cos\theta}+\frac{dJ_y/dt}{\sin\theta\cos\phi\cos\theta}+\frac{dJ_z/dt}{\sin^2\theta\cos\phi\sin\phi}=0.
160: \label{eq:dJ0}
161: \ee
162: From the definition of angular momentum it is also easy to derive
163: \be
164: \calJ_x+\calJ_y+\calJ_z=0,
165: \label{eq:calJ0}
166: \ee
167: where
168: \be
169: \cases{
170: \calJ_x\equiv J_x/(\sin\theta\sin\phi\cos\theta),\cr
171: \calJ_y\equiv J_y/(\sin\theta\cos\phi\cos\theta),\cr
172: \calJ_z\equiv J_z/(\sin^2\theta\cos\phi\sin\phi)
173: \label{eq:calJ}
174: }
175: \ee
176: are determined directly from the position and velocity of the star. Note
177: that equations (\ref{eq:dJ0}) and
178: (\ref{eq:calJ0}) are rotationally invariant, that is, they do not change when
179: arbitrary rotations are applied to their arguments. Below we apply the above analysis to
180: the motion of stars in two simple cases of non-spherical potentials.
181: \begin{itemize}
182: \item If the non-spherical component of the gravitational potential is
183: axisymmetric about the plane $z=0$,
184: \be
185: \Phi\nsph=\Phi\nsph(R=\sqrt{x^2+y^2},z),
186: \label{eq:Phiaxis}
187: \ee
188: then $\partial\Phi\nsph(R,z)/\partial x^2=\partial\Phi\nsph(R,z)/\partial y^2$, and $J_z$
189: is conserved. Stars ejected from the Galactic center on radial orbits move in a plane
190: with
191: \be
192: \calJ_x=-\calJ_y, \qquad \calJ_z=0.
193: \label{eq:calJaxisy}
194: \ee
195: \item If the non-spherical component of the potential is
196: triaxial,
197: \be
198: \Phi\nsph=\Phi\nsph(x^2+y^2/p^2+z^2/q^2),
199: \ee
200: then a triaxiality parameter can be defined as
201: \be
202: T\equiv {p^{-2}-1 \over q^{-2}-1}.
203: \label{eq:Tdef}
204: \ee
205: If $p=q=1$, the potential reduces to the spherical case. If $p=1$ and $q\ne1$ ($T=0$),
206: $q=1$ and $p\ne1$, or $p=q\ne1$ ($T=1$), the potential is axisymmetric.
207: If $q<p<1$, the triaxiality parameter is $0<T<1$. In a triaxial potential,
208: equation (\ref{eq:dJidt}) can be written as
209: \be
210: \frac{dJ_z/dt}{\sin^2\theta\cos\phi\sin\phi}=-T \frac{dJ_y/dt}{\sin\theta\cos\phi\cos\theta}.
211: \label{eq:TdJ}
212: \ee
213: For HVSs moving away from the Galactic center on radial orbits,
214: the deviation of their trajectory from the initial ejection direction,
215: ($\delta\theta,\delta\phi$), is small. Replacing the angles $(\theta,\phi)$ in
216: equation (\ref{eq:TdJ}) with $(\theta\pm\delta\theta,\phi\pm\delta\phi)$ and integrating
217: yields
218: \be
219: T &= &-\frac{\calJ_z}{\calJ_y}[1\mp\frac{2\delta\theta}{\sin(2\theta)}\mp\frac{\delta\phi}{\tan\phi}+\frac{\delta^2\theta}{\sin^2\theta}+\frac{\delta^2\phi}{\sin^2\phi}+ \nonumber \\
220: & & \frac{2\delta\theta\delta\phi}{\sin(2\theta)\tan\phi}+...],
221: \label{eq:TcalJ}
222: \ee
223: where the $\sin$ and $\cos$ arguments have been kept constant in the integration.
224: The term in parenthesis specifies the maximum systematic error on the triaxiality
225: estimator $T=-\calJ_z/\calJ_y$, and numerical calculations (see \S\ 4 below) of the
226: motion of HVSs in a triaxial potential show that the typical error on $T$ is actually
227: smaller. For a sample of $N$
228: HVSs we can use equation (\ref{eq:TcalJ}) to construct a statistical estimator
229: of triaxiality as
230: \be
231: \bar{T}=\sum_{i=1}^{N}\frac{T_i}{\sigma^2_{T_i}}/\sum_{i=1}^{N}\frac{1}{\sigma^2_{T_i}},
232: \label{eq:Tbar}
233: \ee
234: with standard deviation
235: \be
236: \sigma_{\bar{T}}=\left(\sum_{i=1}^{N}\frac{1}{\sigma^2_{T_i}}\right)^{-1/2}.
237: \label{eq:Terror}
238: \ee
239: Here
240: \be
241: T_i\equiv-\frac{\calJ_{z,i}}{\calJ_{y,i}}=-\frac{J_{z,i}\cos\theta_i}{J_{y,i}\sin\theta_i\sin\phi_i},
242: \ee
243: and $\sigma_{T_i}$ is its error. Note that the parameters $T_i$ are fully determined by the
244: instantaneous positions and velocity vectors of the HVSs in the sample, and that, while
245: the ratio $-\calJ_{z,i}/\calJ_{x,i}$ can also be used to estimate the triaxiality
246: $(1-p^{-2})/(q^{-2}-p^{-2})$, this does not provide any independent
247: information since it can be trivially derived from $\calJ_{z,i}/\calJ_{y,i}$
248: using equation (\ref{eq:calJ0}).
249:
250: In the following we set the $z$-axis of the triaxial potential to be normal
251: to the Galactic disk plane, and denote with $\eta_0$ the angle measured
252: counter-clockwise from a reference direction (e.g. the line from the Galactic
253: center to the Sun) to the $x$-axis of the potential. The ratio $-\calJ_{z,i}'/\calJ_{y,i}'$
254: in a frame forming an angle $\eta$ with the reference direction can be written as
255: \be
256: -\frac{\calJ_{z,i}'}{\calJ_{y,i}'}=\frac{\frac{1}{p^2}-1}{\left(\frac{1}{q^2}-1\right)A-\left(\frac{1}{p^2}-1\right)B},
257: \ee
258: where
259: \be
260: A=\frac{\sin(2\phi_i')}{\sin[2(\eta-\eta_0+\phi_i')]},\qquad
261: B=\frac{\sin(\eta-\eta_0)\sin\phi_i'}{\cos(\eta-\eta_0+\phi_i')}.
262: \ee
263: If $\eta\ne\eta_0$, the values $-\calJ_{z,i}'/\calJ_{y,i}'$ may spread out over a large
264: range due to the different angles $\phi_i'$ in the sample. The angle $\eta_0$ can
265: be estimated by minimizing the weighted variance of
266: $T_i'\equiv -\calJ_{z,i}'/\calJ_{y,i}'$:
267: \be
268: \sum_{i=1}^{N}\left(\frac{T_i'-\bar{T'}}{\sigma^2_{T_i'}/\sigma^2_{\bar{T}'}}\right)^2,
269: \label{eq:variance}
270: \ee
271: where $\sigma_{T_i'}$ is the error of $T_i'$ and the values of $\bar{T'}$ and
272: $\sigma^2_{\bar{T}'}$ are obtained from equations (\ref{eq:Tbar}) and
273: (\ref{eq:Terror}).
274: \end{itemize}
275:
276: \section{Galactic gravitational potential}\label{sec:potential}
277:
278: We use here a four-component model for the gravitational potential of the Milky Way,
279: $\Phi=\Phi\bh+\Phi\bulge+\Phi\disk+\Phi\halo$, where (cf. \citealt{Gnedin05}):
280: %
281: \begin{itemize}
282:
283: \item $\Phi\bh$ is the contribution of Sgr A$^*$,
284: \be
285: \Phi\bh=-\frac{GM\bh}{r},
286: \label{eq:Phibh}
287: \ee
288: with mass $M\bh\simeq 3.6\times 10^6\msun$ \citep{Ghez05,Eisenhauer05}.
289: The radius of influence of Sgr A$^*$ is $GM\bh/\sigma_c^2\simeq 1.6\pc\
290: (M\bh/3.6\times10^6\msun) (100\kms/\sigma_c)^2$,
291: where $\sigma_c$ is the one-dimensional stellar velocity dispersion in the
292: Galactic center.
293:
294: \item $\Phi\bulge$ is the contribution of the spherical bulge \citep{Hern90},
295: \be
296: \Phi\bulge=-\frac{GM\bulge}{r+a\bulge},
297: \label{eq:Phib}
298: \ee
299: with mass $M\bulge=10^{10}\msun$ and core radius $a\bulge=0.6\kpc$. As both
300: the bulge mass and size are small compared to those of the
301: Galactic disk and halo, a slight deviation of the bulge from
302: sphericity will not have a significant effect on the change of angular
303: momentum of HVSs in the halo.
304:
305: \item $\Phi\disk$ is the contribution of the axisymmetric disk \citep{MN75},
306: \be
307: \Phi\disk(R,z)=-\frac{GM\disk}{\sqrt{R^2+\left(a\disk+\sqrt{z^2+b\disk^2}\right)^2}},
308: \label{eq:Phid}
309: \ee
310: with mass $M\disk=4\times 10^{10}\msun$, scale length $a\disk=5\kpc$, and
311: scale height $b\disk=0.3\kpc$.
312:
313: \item $\Phi\halo$ is the contribution of the triaxial dark matter halo,
314: \begin{eqnarray}
315: \Phi\halo(x,y,z)& = &\Phi_{\rm NFW}(r^t),\\
316: r^t& = &p^{1/3}q^{1/3}\left(x^2+\frac{y^2}{p^2}+\frac{z^2}{q^2}\right)^{1/2}, \label{eq:Phihpq}\\
317: \Phi_{\rm NFW}(r^t)& = &-\frac{GM_{200}}{r_sf(c)}\frac{\ln(1+r^t/r_s)}{r^t/r_s}.
318: \end{eqnarray}
319: Here $c\equiv r_{200}/r_s$ is the halo concentration parameter, $r_{200}$ the radius
320: within which the enclosed average density is 200 times the mean matter density,
321: $r_s$ the scale radius, and $f(c)=\ln(1+c)-c/(1+c)$. This generalization of an NFW
322: \citep{NFW96} model ensures that
323: the spherically-averaged potential of the triaxial halo is similar to that of
324: a spherical halo with the same mass $M_{200}$ and scale radius \citep{HNS06}.
325: We choose the following parameters for the Milky Way halo: $c=15$,
326: $r_{200}=389\kpc$, $M_{200}=1.8\times10^{12}\msun$ (e.g. \citealt{Diemand07}), and assume
327: that the $z$-axis of the halo and disk potentials coincide.
328: We also assume, for simplicity, that the asphericity of the potential is
329: constant with radius $r$, and set $p=0.8$ and $q=0.7$ in the calculations below.
330:
331: \end{itemize}
332:
333: CDM halos are not smooth but have a wealth of substructure on all resolved mass
334: scales (e.g. \citealt{Moore99,Klypin99}). The assumption made above of a
335: smooth gravitational potential neglects the deflection of HVS trajectories by halo
336: substructure. A star with velocity $v$ passing within a distance
337: $r_* $ from a subhalo of mass $M\sub$ will change its velocity
338: by an amount $\delta v=2GM\sub/(vr_*)$. The mass function of substructure in
339: a Milky Way-sized halo can be described as $N(>M\sub) =6.4\times 10^{-3}
340: (M_{200}/M\sub)$ in the subhalo mass range
341: $10^6\msun<M\sub<f_{\rm max}M_{200}$ \citep{Diemand07}, with $f_{\rm max}=0.01$.
342: The probability that a HVS ejected from the Galactic center is subject to a
343: velocity deflection $>\delta v$ can then be estimated as
344: \begin{eqnarray}
345: P(>\delta v)&=&{3\over 4}\int \frac{r^2_*}{r_{200}^2}\,\frac{dN}{dM\sub}
346: dM\sub \nonumber \\
347: &\sim& 0.08 \left(\frac{f_{\rm max}}{0.01}\right)\,\left(\frac{V_c}{140\,\kms}\right)^4
348: \times \nonumber \\
349: & & \left(\frac{10^3\kms}{v}\right)^2\left(\frac{1\kms}{\delta v}\right)^2,
350: \end{eqnarray}
351: where $V_c\equiv (GM_{200}/r_{200})^{1/2}$ is the halo circular velocity. This
352: probability is quite low so that a smooth potential is a good assumption in
353: this work.
354:
355: \section{Numerical calculations}\label{sec:simulation}
356:
357: In this section we perform numerical calculations of the motion of HVSs
358: in the MW gravitational potential.
359:
360: \begin{figure*}
361: \begin{center}
362: %\epsscale{0.8} \plotone{map.epsi}
363: %\includegraphics[width=0.85\textwidth,angle=0]{map.epsi}
364: \includegraphics[width=0.85\textwidth,angle=0]{mapbitmap.ps}
365: \caption{Present-day spatial distribution and velocity vectors of HVSs ejected by a binary
366: BH at the Galactic center. The length of each vector is proportional to speed.
367: Ten thousand HVSs were generated at a constant rate in the past $10^9\yr$, with a
368: velocity and spatial distribution obtained from the three-body scattering experiments
369: of \citet{SHM06}. The binary has a mass $3.6\times 10^6\msun$, mass ratio of 1/81, orbital
370: semimajor axis equal to $0.1a\h$, and eccentricity 0.3, and orbits in the
371: $(x,y)$ plane. The reference axes are set along the $(x,y,z)$-axes of
372: the triaxial halo potential in eq. (\ref{eq:Phihpq}).
373: } \label{fig:map} \end{center} \end{figure*}
374:
375: \begin{figure*}
376: \begin{center}
377: %\epsscale{0.8} \plotone{alphar.epsi}
378: \includegraphics[width=0.9\textwidth,angle=0]{alphar.epsi}
379: \caption{Panels (a)-(c): Deflection angle $\alpha=\arcsin(|\vec{r}\times\vec{v}|/rv)$
380: between velocity and position vectors for all HVSs plotted in Fig. \ref{fig:map},
381: as a function of Galactocentric distance $r$. The panels show different velocity
382: ranges or different scales in $\alpha$. Panel (d): transverse velocity in the Galactocentric
383: frame versus distance $r$ for all stars shown in Panel (c). {\it Panels (a)--(d)}:
384: triaxial halo potential with $p=0.8$ and $q=0.7$ (see eq. \ref{eq:Phihpq}).
385: {\it Panels (e)--(h)}: spherical halo potential with $p=q=1$.
386: Different colors depict different velocity ranges: 300-400$\kms$
387: ({\it magenta}), 400--600$\kms$ ({\it red}), 600--800$\kms$ ({\it blue}), 800--1000$\kms$
388: ({\it green}), 1000--1200$\kms$ ({\it cyan}), $>1200\kms$ ({\it black}).
389: HVSs with angles between $5^\circ$ and $180^\circ$
390: in Panels (a), (b), (e) and (f) are bound stars with significantly bent orbits, and their
391: detailed $\alpha$ distribution depends on halo triaxiality. Stars with small deflections
392: ($\alpha \la5^\circ$) are either unbound (and have distances $\ga 10\kpc$) or
393: are in the initial phases of their orbital period (and are closer to the Galactic center, see
394: also Fig. \ref{fig:alphat}). The large cross in Panel (a) flags the locus of
395: the bound star whose orbit is shown in Fig. \ref{fig:orbit}.
396: } \label{fig:alphar} \end{center} \end{figure*}
397:
398:
399: \subsection{Initial conditions}\label{subsec:initial}
400:
401: According to the study of \citet{YT03}, three-body interactions between
402: ambient stars and a BH pair (where the secondary BH may be an IMBH
403: inspiralling towards Sgr A$^*$; e.g., \citealt{BGPZ06}) expel HVSs
404: ($v>10^3\kms$) at a rate that can be as large as $\sim 10^{-4}\yr^{-1}$
405: (for a binary with semimajor axis $0.5\times 10^{-3}\pc$ and mass ratio of 0.01).
406: Tidal break-up of binary stars (``Hills' mechanism'') ejects
407: HVSs at a rate $\sim 10^{-5}(\eta/0.1)\yr^{-1}$,
408: where $\eta$ is the fraction of stars in binaries with semimajor axis $\la
409: 0.3\,\au$. Close encounters of two single stars in the vicinity of Sgr A$^*$ may
410: also produce HVSs but at the negligible rate of $10^{-11}\yr^{-1}$.
411: Different ejection mechanisms give origin to different spatial and velocity
412: distributions of HVSs. While Hills' mechanism predict HVSs to be expelled
413: isotropically at an approximately constant rate, in models involving a BH pair
414: HVSs are ejected preferentially within the orbital plane of the binary
415: in a short burst lasting a few Myr \citep{ZB01,Levin06,SHM06,SHM07}. In the latter
416: case the degree of anisotropy depends on binary separation, the mass ratio,
417: and the orbital eccentricity of the BH binary.
418:
419: For illustrative purposes, we assume in this paper that HVSs are ejected from
420: the Galactic center by the BH binary mechanism. We use the stellar spatial and
421: velocity distributions derived from the scattering experiments of \citet{SHM06} for a binary
422: with mass ratio $M_2/M_1=1/81$, semimajor axis $a=0.1\,a_h$, and eccentricity $e=0.3$.
423: The binary orbit is in the Galactic disk plane, and the velocity of the lighter hole at
424: pericenter is directed along $\phi=3\pi/2$. The ``hardening'' radius $a_h$ is defined as
425: \citep{Q96}
426: \begin{eqnarray}
427: & & a_h \equiv \frac{GM_2}{4\sigma_c^2} \nonumber \\
428: &\simeq& 0.39\pc\left(\frac{M_2}{M_1}\right)
429: \left(\frac{M_1}{3.6\times10^6\msun}\right)
430: \left(\frac{100\kms}{\sigma_c}\right)^2.
431: \label{eq:ah}
432: \end{eqnarray}
433: When $a>a_h$, the binary
434: separation decreases both by dynamical friction and three-body interactions with low-angular
435: momentum stars passing in its immediate vicinity. After the binary becomes ``hard'' ($a<a\h$),
436: the bound pair loses orbital energy mainly through three-body interactions until gravitational radiation takes over \citep{BBR80,Y02}.
437: The ejection speed of the stars at infinity spans a range with r.m.s. $\sim
438: 7\times10^2\kms(M_2/0.01M\bh)^{1/2}(10^{-3}\pc/a)^{1/2}(M\bh/3.6\times10^6\msun)^{1/2}$ (see eq.~33 in \citealt{YT03}). In the calculations below, we assume $10^4$
439: HVSs are ejected from the Galactic center at a constant rate over the last $10^9$ years,
440: and ignore for simplicity the orbital evolution of the binary during such timescale.
441: Different stars move independently in the Galactic potential. A
442: fourth-order Runge-Kutta method with adaptive stepsize control is used
443: to solve the differential equations of the motion of the stars.
444: Note that the
445: adopted mass ratio and semimajor axis are within the allowed parameter space for a
446: BH pair at the Galactic center (see Fig. 2 in \citealt{YT03}).
447:
448: \subsection{Results} \label{subsec:results}
449:
450: \begin{figure}
451: \begin{center}
452: %\epsscale{0.8} \plotone{orbit.epsi}
453: %\includegraphics[width=0.8\textwidth,angle=0]{orbit.epsi}
454: \includegraphics[width=\hsize,angle=0]{orbit.epsi}
455: \caption{The bound orbit of a simulated HVS ejected from the Galactic center. The
456: orbit is significantly bent, rather than radial as expected in a spherical potential.
457: The star's travel time $820\Myr$, and its present-day velocity and Galactocentric distance
458: are $300\kms$ and $10\kpc$, respectively.
459: }
460: \label{fig:orbit} \end{center} \end{figure}
461:
462: \begin{figure}
463: \begin{center}
464: %\epsscale{0.6} \plotone{histogram.epsi}
465: %\includegraphics[width=0.6\textwidth,angle=0]{histogram.epsi}
466: \includegraphics[width=\hsize,angle=0]{histogram.epsi}
467: \caption{Histogram of the deflection angle distribution of all stars shown in
468: Fig.\ \ref{fig:alphar}(a) ({\it solid line}) and (e) ({\it dotted line}) with
469: $r<20\kpc$ and $300\kms<v<400\kms$. The bound stars are clustered around
470: $0^\circ$ and $180^\circ$ in a spherical halo potential ({\it dotted line}),
471: while their distribution is spread uniformly over the range $10^\circ$--$180^\circ$
472: in a triaxial halo potential ({\it solid line}).}
473: \label{fig:histogram} \end{center} \end{figure}
474:
475: \begin{figure}
476: \begin{center}
477: %\epsscale{0.6} \plotone{alphat.epsi}
478: %\includegraphics[width=0.6\textwidth,angle=0]{alphat.epsi}
479: \includegraphics[width=\hsize,angle=0]{alphat.epsi}
480: \caption{Deflection angle versus travel time from the Galactic center.
481: {\it Top panel:} all stars shown in Fig. \ref{fig:alphar}(a) with $r<50\kpc$ and
482: $300\kms<v<400\kms$. {\it Bottom panel:} all stars shown in Fig.
483: \ref{fig:alphar}(b) with $r<200\kpc$ and $400\kms<v<600\kms$. HVSs
484: with short travel times, $\la 50$ Myr ({\it top}) and $\la 400\Myr$ ({\it bottom})
485: have $\alpha$ angles smaller than $5^\circ$. The large cross in the top panel
486: flags the locus of the bound star whose orbit is
487: shown in Fig. \ref{fig:orbit}.
488: }
489: \label{fig:alphat} \end{center} \end{figure}
490:
491: We use the initial conditions described above and numerically integrate the orbits of
492: HVSs in the Galactic potential. Figures \ref{fig:map} and \ref{fig:alphar} show maps of
493: stellar position and velocity vectors at the present time and their deflection angles
494: $\alpha=\arcsin(|\vec{r}\times\vec{v}|/rv)$. For HVSs with $v\ga 600\,\kms$, deviations
495: are quite small, $\alpha\la 5^\circ$, at all distances within $200\,\kpc$. Lower velocity
496: stars at small distances can instead be bound, and their deflection angles extend to
497: $180^\circ$. Many of the stars with $300\kms\la v\la 400\kms$ and $r\la 20\,\kpc$ in
498: Figure \ref{fig:alphar}(a) follow bound trajectories,
499: while no stars in this velocity range have substantial deviation angles at large
500: Galactocentric distances. One example of a bound orbit is shown in
501: Figure \ref{fig:orbit}, where the trajectory has been significantly bent by the
502: triaxial halo and the flattened disk potentials, and the star does not return to the Galactic center.
503: Note that bound stars with $v>400\,\kms$ and large $\alpha$'s are typically
504: observed at smaller distances ($r\la 10\,\kpc$) than bound stars of lower velocities.
505: This is because a larger ejection speed from the Galactic center is needed to
506: maintain a high velocity at large distances, and such stars may then have either escaped
507: from the halo or not have had sufficient time to come back and show a significant
508: bend in their orbits. (For reference, the local escape speed is about
509: $500-600\,\kms$, \citealt{Smith06}.) We find that 60\% of the stars with $v>300\kms$ at
510: $r<20\kpc$ have velocities $v<400\kms$, and 20\% of the stars with $v>300\kms$ at
511: $r<10\kpc$ have velocities $v<600\kms$.
512:
513: \begin{figure*}
514: \begin{center}
515: %\epsscale{0.8} \plotone{mapJn.epsi}
516: \includegraphics[width=0.9\textwidth,angle=0]{mapJn.epsi}
517: \caption{The variables $-\calJ_z/\calJ_y$ and $(v/r)\calJ_y$ versus
518: Galactocentric distances $r$ for all HVSs far from the plane of the
519: disk, $|\theta-90^\circ|<30^\circ$. Stars near the plane are
520: included only in Panel (d). Different colors represent different velocity
521: ranges as in Fig. \ref{fig:alphar}. Panel (c) shows $\calJ_y$ on a different
522: scale than Panel (b), and the scattered magenta
523: dots are bound stars with significantly bent orbits. Panel (e) shows the
524: effect of removing the disk potential from the calculation. Panel (f) shows
525: the results in a reference frame with $\eta-\eta_0=10^\circ$, different from
526: the triaxial frame of the halo potential. The solid line represents the
527: triaxiality of the halo potential $T=0.54$ (with $p=0.8$ and $q=0.7$ in
528: eq.~\ref{eq:Tdef}) assumed in the calculation.
529: } \label{fig:mapJn} \end{center} \end{figure*}
530:
531:
532: Figure \ref{fig:alphat} shows that bound stars with large deflection angles
533: ($5^\circ\la\alpha\la180^\circ)$ have generally traveled a long time after
534: ejection ($\ga$50 Myr for $r\la 20\,\kpc$), and many of them have experienced at least one
535: orbital period (Fig. \ref{fig:orbit}). HVSs with small deflections ($\alpha\la5^\circ$)
536: have instead a short travel time and are generally on the initial phases of their
537: first orbital periods (see the concentration of the stars at the left bottom of the
538: panels in Figure \ref{fig:alphat}). The transverse velocities (in the Galactocentric
539: frame) of stars with $v>300\kms$ are typically higher than $3\,\kms$, and can be up to
540: $30\,\kms$ (hundreds of $\kms$) for unbound (bound) HVSs. Note that
541: $3\,\kms$ corresponds at a distance of 100 kpc to a proper motion of $20\,\mu$as in
542: three years, which can be resolved by the next generation of astrometric
543: surveys like GAIA. According to our calculations most HVSs would have transverse
544: velocities $\ga 3\,\kms$ even in the case of a weakly triaxial halo with
545: $(p,q)=(0.95,0.9)$: these transverse velocities are larger than those
546: associated with the bending of stellar trajectories caused by the axisymmetric disk
547: (see Fig.\ \ref{fig:alphar}h).
548:
549: A comparison between Figures \ref{fig:alphar}(e)--(g) and \ref{fig:alphar}(a)--(c)
550: shows that the distribution of deflection angles versus distance is different
551: for trajectories in a spherical rather than triaxial halo. In
552: Figure \ref{fig:alphar}(e) and (f), the $\alpha$ angles of bound stars are
553: clustered around $0^\circ$ and $180^\circ$ because their orbits are
554: highly eccentric with little bending (see also Fig.~\ref{fig:histogram}). By contrast, in Figure \ref{fig:alphar}(a) and
555: (b) the deflection angles of bound stars lies at intermediate values.
556: This difference can be used as an indicator of the triaxiality of the MW
557: dark matter halo.
558:
559: \begin{figure*}
560: \begin{center}
561: %\epsscale{0.8} \plotone{mapJns.epsi}
562: \includegraphics[width=0.8\textwidth,angle=0]{mapJns.epsi}
563: \caption{The variable $(v/r)\calJ_z$ versus $(v/r)\calJ_y$. The solid and
564: dashed lines correspond to $\calJ_z=-T\calJ_y$ and $\calJ_z=0$ (see
565: eq.~\ref{eq:calJaxisy}), respectively. The stars are those shown in Fig.
566: \ref{fig:mapJn}. Panel (a) includes only stars far from the plane of the
567: disk, $|\theta-90^\circ|>30^\circ$, while Panel (b) includes all stars.
568: The effect of the disk potential can be gauged from the spread of the stellar
569: dots from the solid line to the dashed line in Panel (b). The magenta and red dots
570: scattered above the dashed line or below the solid line represent bound stars with
571: bent orbits.
572: } \label{fig:mapJns} \end{center} \end{figure*}
573:
574:
575: Figure \ref{fig:mapJn}(a)--(c) depicts the values of $-\calJ_z/\calJ_y$ and
576: $(v/r)\calJ_y$ versus distance of all the HVSs plotted in Figure \ref{fig:map}
577: having $v>300\kms$ and far from the plane of the disk, i.e. with $|\theta-90^\circ|>30^\circ$.
578: Figure \ref{fig:mapJn}(d) shows all stars with $\theta$ in the range
579: $0^\circ$--$180^\circ$. As seen in Figure \ref{fig:mapJn}(a), the ratio
580: $-\calJ_z/\calJ_y$ at large distances ($r\ga 100\kpc$) is close to the
581: triaxiality of the halo potential ($T=0.54$, {\it solid line}) assumed in our
582: calculations, with only a small scatter. The scatter is larger in the quantity
583: $(v/r)\calJ_y$ plotted in Figure \ref{fig:mapJn}(b). The slight offset ($\la10\%$)
584: of the stellar dots from the solid line is partly due to the disk
585: potential that causes an additional bending of stellar trajectories towards the disk
586: plane, increasing $|\calJ_y|$ without changing $\calJ_z$ (see the Appendix for details about
587: the correction of such offset owing to the Galactic disk). As shown in Figure \ref{fig:mapJn}(e),
588: HVSs become better tracers of triaxiality after removing the disk potential from our
589: calculations. The higher the velocities, the smaller the offset. The curvature of the
590: ratio $-\calJ_z/\calJ_y$ towards smaller values at small distances is also
591: an effect of the disk. The ratio $-\calJ_z/\calJ_y$ for stars close to disk plane is not a good
592: approximation of halo triaxiality even at large distances (see Fig.\ \ref{fig:mapJn}d).
593: We plot the values of $\calJ_z$ versus $\calJ_y$ in Figure \ref{fig:mapJns}, where Panel
594: (a) shows stars with $|\theta-90^\circ|<30^\circ$, and Panel (b) stars with $\theta$ in
595: the full range $0^\circ$--$180^\circ$. The effect of the disk potential can be seen
596: in the spreading of the stellar dots from the solid line $\calJ_z=-T\calJ_y$ to the
597: dashed line $\calJ_z=0$.
598: Note that the scatter in the ratio $-\calJ_z'/\calJ_y'$ increases
599: significantly if the reference frame is different from the frame of the triaxial halo
600: potential (Fig. \ref{fig:mapJn}f).
601:
602:
603: \section{Summary}\label{sec:conclusion}
604:
605:
606: We have studied the unique kinematics of HVSs ejected from the
607: Galactic center with almost zero initial specific angular momentum.
608: HVSs can travel in the Galactic halo on either bound or unbound orbits,
609: and their spatial and velocity distribution
610: at large Galactocentric distances ($r\ga50\kpc$) contain information
611: on the asphericity of the halo gravitational potential. We have
612: proposed an estimator of the triaxiality of the Galactic dark matter
613: halo that is determined solely by instantaneous position and velocity
614: vectors of HVSs, is independent of the details of the ejection mechanism,
615: and does not require an accurate knowledge of halo mass.
616: Future astrometric and deep wide-field surveys of HVSs should detect significant numbers
617: of HVSs, which could be used to determine the triaxiality of the
618: MW halo by applying the method proposed in this paper.
619:
620: The new class of possibly bound HVSs with velocities $+275<v_{\rm rf}<+450\,\kms$
621: recently observed by \citet{Brown07} has Galactocentric distances in the range
622: 30-60$\kpc$ or 10-20$\kpc$ depending on whether they are main-sequence or blue
623: horizontal branch stars. In the first case (main-sequence stars at large distances),
624: they have a lifetime of $\la 100\Myr$ and are, according to Figure \ref{fig:alphat},
625: on the initial phases of their first orbital periods. Their deflection
626: angles are expected to be rather small, supporting the fact that a significant
627: excess of B-type stars is observed only at large positive velocities \citep{Brown07}.
628: If bound HVSs are blue horizontal branch stars instead at smaller distances,
629: the travel time of the stars can be much longer than $50\Myr$ as at
630: the ejection moment the stars may not necessarily be blue horizontal
631: branch stars but at some pre-blue-horizontal-branch stage.
632: Such stars may have experienced at least one orbital period, and many of them should
633: be returning to the Galactic center or their orbits should have been significantly
634: bent by the asymmetric Galactic potential (Fig. \ref{fig:alphat}). This scenario
635: appear unlikely since it does not agree with the observed positive radial velocities.
636:
637: It is interesting at this stage to provide an example of a statistical estimate of
638: halo triaxiality from a mock sample. Let us assign a measurement error of
639: $\sigma_a=3\,\kms$ ($a=x,y,z$)
640: to the one-dimensional velocities of all stars in Figure \ref{fig:mapJn}(a)
641: having $v>300\kms$, $|\theta-90^\circ|>30^\circ$, and $55\,\kpc<r<200\,\kpc$
642: (our error analysis assumes that the distances to HVSs are known to within
643: ten percent).
644: We have simulated the observed velocities of such a sample, and plot in
645: the top panel of Figure \ref{fig:Texample} the values of
646: $T_i$ and $\sigma_{T_{i}}$ derived for each HVS. Using equations (\ref{eq:Tbar}) and
647: (\ref{eq:Terror}), we obtain $\bar{T}=0.50$ and $\sigma_{\bar{T}}=0.02$.
648: For $\sigma_a=5\,\kms$, we obtain $\bar{T}=0.50$ and $\sigma_{\bar{T}}=0.03$.
649: These errors are comparable with the systematic error caused by the
650: flattened disk (see the slight offset of dots from the solid line in Fig.
651: \ref{fig:mapJn}a). We have tried different values of the gravitational potential
652: parameters $(p,q)$ in the calculations, and found that the estimated value of $\bar{T}$
653: is always consistent with the assumed halo triaxiality (see middle and bottom panels in Fig.
654: \ref{fig:Texample}).
655:
656: \begin{figure}
657: \begin{center}
658: %\epsscale{0.8} \plotone{Texample.epsi}
659: %\includegraphics[width=0.8\textwidth,angle=0]{Texample.epsi}
660: \includegraphics[width=\hsize,angle=0]{Texample.epsi}
661: \caption{Simulated triaxiality parameters and their errors from a mock
662: sample of HVSs with $v>300\kms$, $|\theta-90^\circ|>30^\circ$, and $55\kpc<r<200\kpc$.
663: An observational error of $3\kms$ is assumed for in ``measured'' one-dimensional velocity.
664: From top to bottom, the values of $(p,q)$ used in the calculations
665: are (0.8, 0.7), (0.8, 0.6), and (0.8, 0.75), corresponding to a triaxiality
666: $T$=0.54, 0.32, and 0.72, respectively ({\it horizontal lines}).
667: A statistical analysis of the sample using eqs.
668: (\ref{eq:Tbar}) and (\ref{eq:Terror}) yields $(\bar{T},\sigma_{\bar{T}})$=(0.50,
669: 0.02), (0.31, 0.01), and (0.64, 0.02), respectively. For clarity, only 1/3
670: of the sample points are drawn in the figure.}
671: \label{fig:Texample} \end{center} \end{figure}
672:
673:
674: \begin{figure}
675: \begin{center}
676: %\epsscale{0.8} \plotone{variance.epsi}
677: %\includegraphics[width=0.8\textwidth,angle=0]{variance.epsi}
678: \includegraphics[width=\hsize,angle=0]{variance.epsi}
679: \caption{Variance of the triaxiality parameter $T'$ of the simulated sample (top panel
680: in Fig. \ref{fig:Texample}) in different reference frames as a function of
681: $\eta-\eta_0$. Here $\eta_0$ is the angle measured
682: counter-clockwise from a reference direction (e.g. the line from the Galactic
683: center to the Sun) to the $x$-axis of the halo potential, and
684: $\eta$ is the angle formed by the observational frame with the reference direction.
685: The variance has a minimum at $\eta-\eta_0=0^\circ$.} \label{fig:variance} \end{center} \end{figure}
686:
687:
688: If the $x$-axis of the halo potential form an angle $\eta_0$ (measured counter-clockwise)
689: with the reference axis of the observations, then (as seen from Fig. \ref{fig:variance})
690: the minimum variance of $\calJ_{z,i}'/\calJ_{y,i}'$ in a set of simulated samples
691: each in a frame at angle $\eta$ from the observational reference axis
692: (see eq.~\ref{eq:variance}) occurs for $\eta=\eta_0$. Our calculations show that
693: an error of $5^\circ$ in the estimate of $\eta_0$ may cause an error of
694: 2\% in the estimate of $\bar{T}$.
695:
696: Note that the axis ratios $(p,q)$ of the halo potential are
697: degenerate in the defined triaxiality parameter (eq.~\ref{eq:Tdef}). After
698: determining $T$, the values of $(p,q)$ could be also obtained by using any
699: value of $(\calJ_x, \calJ_y,\calJ_z)$ (e.g., see Fig.~\ref{fig:mapJn}b), but
700: the modeling would be sensitive to halo mass and the shape of halo potential
701: used (see also the determination of axis ratios in \citealt{Gnedin05}
702: by tracing back HVS orbits).
703:
704: Finally, it is possible that a few HVSs in the halo may be
705: produced by the interactions of stars with an IMBH in satellite galaxies like
706: the Large Magellanic Cloud \citep{Edelmann05,GPZ07}, and that these
707: would contaminate the sample ejected from the Galactic center. The ejection
708: rates from the satellite dwarfs of the MW are expected to be much smaller than the rate from
709: Sgr A$^*$, however. Such ``satellite'' HVSs will also have much larger angular
710: momenta in the Galactocentric rest-frame, and should be easily distinguishable
711: from the Galactic center sample.
712:
713: %\acknowledgements
714:
715: We have benefited from discussions with Francesco Haardt, Youjun Lu, and Scott
716: Tremaine. We thank Alberto Sesana for providing the initial velocity and spatial
717: distributions of HVSs ejected from a black hole binary. P.M. acknowledges
718: financial support from NASA through grants NAG5-11513 and NNG04GK85G, and
719: from the Alexander von Humboldt Foundation. Q.Y. acknowledges initial support
720: from NASA through Hubble Fellowship grant HST-HF-01169.01-A awarded by the Space Telescope
721: Science Institute, which is operated by the Association of Universities for
722: Research in Astronomy, Inc., for NASA, under contract NAS 5-26555.
723:
724: \appendix
725:
726: \section{Correction of the effect of disk potential on the inferred
727: halo triaxiality}
728:
729: As shown in Figure \ref{fig:mapJn}(a), the triaxiality obtained through
730: $T=-\calJ_z/\calJ_y$ has a small offset from the true halo triaxiality. Part of
731: this offset comes from the effects of the flattened disk potential, and part from
732: the fixing of $(\theta,\phi)$ in the integration of equation (\ref{eq:TdJ}).
733: The offset due to the constant $(\theta,\phi)$ assumption is smaller for stars
734: of higher velocities, as seen in Figure \ref{fig:mapJn}(e) that shows the
735: results after the disk potential was removed. Below we provide a method to correct
736: the offset due to the Galactic disk. As in the determination of halo triaxiality
737: proposed in this work, even for this correction it is not necessary to trace back the
738: orbits of HVSs.
739:
740: The contribution to the change in specific angular momentum due to the disk
741: potential can be expressed as:
742: \be
743: J_{y,}{\disk}=\int^{r_0}_{r_{\rm ej}}\frac{dJ_{y,}{\disk}}{dt}\frac{dt}{dr}dr,
744: \ee
745: where
746: \be
747: \frac{dJ_{y,}{\disk}}{dt}=2xz\left(\frac{\partial\Phi\disk}{\partial z^2}-\frac{\partial\Phi\disk}{\partial x^2}\right),
748: \label{eq:dJdisk}
749: \ee
750: \be
751: \frac{dt}{dr}=\frac{1}{v_r}\simeq \frac{1}{\sqrt{v^2_0+2{\Phi}(\vec{r}_0)-2{\Phi}(\vec{r})}},
752: \label{eq:vr}
753: \ee
754: $r_{\rm ej}$ is the initial distance of the HVS from the Galactic center at
755: ejection, $\vec{r}_0$ and $v_0$ are the current position and velocity of the
756: HVS, $r_0=|\vec{r}_0|$, and $v_r$ is its radial velocity at $r$.
757: %In equation (\ref{eq:vr}), the radial velocity component dominates.
758: We can remove the effect of the disk potential from the variable $\calJ_y$ by
759: computing $\calJ_y-\calJ_{y,}{\disk}$, where
760: \begin{eqnarray}
761: & & \calJ_{y,}{\disk} = \frac{J_{y,}{\disk}}{\sin\theta\cos\phi\cos\theta}\nonumber \\
762: & = & \int^{r_0}_{r_{\rm ej}}\frac{G M\disk}{(r^2+a\disk^2+b\disk^2+2a\disk
763: \sqrt{r^2\cos^2\theta+b\disk^2})^{3/2}} \nonumber\\
764: & & \frac{a\disk r^2}
765: {\sqrt{r^2\cos^2\theta+b\disk^2}}\frac{dr}{v_r},
766: \label{eq:calJydisk}
767: \end{eqnarray}
768: and where equation (\ref{eq:Phid}) describing the disk
769: potential has been used. The angles $(\theta,\phi)$ are fixed in the integration
770: to be present-day values. In addition to the disk potential, we also need to assume a
771: halo potential to determine $v_r$, for which we may use the spherical part of
772: the halo potential by setting $(p,q)=(1,1)$.
773: We have tested this correction and found that half of the offset to the true
774: halo triaxiality can be corrected.
775:
776: \begin{thebibliography}{}
777:
778: \bibitem[Baumgardt et al.(2006)]{BGPZ06} Baumgardt, H., Gualandris, A., \& Portegies Zwart, S.\ 2006, MNRAS, 372, 174
779:
780: \bibitem[Begelman, Blandford \& Rees(1980)]{BBR80} Begelman, M.\ C., Blandford, R.\ D., \& Rees, M.\ J.\ 1980, Nature, 287, 307
781:
782: \bibitem[Bromley et al.(2006)]{Bromley06} Bromley, B.\ C., Kenyon, S.\ J., Geller, M.\ J., Barcikowski, E.\ B., Warren, R., \& Kurtz, M.\ J.\ 2006, ApJ, 653, 1194
783:
784: \bibitem[Brown et al.(2005)]{Brown05} Brown, W.\ R., Geller, M.\ J., Kenyon, S.\ J., \& Kurtz, M.\ J.\ 2005, ApJ, 622, L33
785:
786: \bibitem[Brown et al.(2006a)]{Brown06a} Brown, W.\ R., Geller, M.\ J., Kenyon, S.\ J., \& Kurtz, M.\ J.\ 2006a, ApJ, 640, L35
787:
788: \bibitem[Brown et al.(2006b)]{Brown06b} Brown, W.\ R., Geller, M.\ J., Kenyon, S.\ J., \& Kurtz, M.\ J.\ 2006b, ApJ, 647, 303
789:
790: \bibitem[Brown et al.(2007)]{Brown07} Brown, W.\ R., Geller, M.\ J., Kenyon, S.\ J., Kurtz, M.\ J., \& Bromley, B.\ C.\ 2007, ApJ, in press (astro-ph/0701600)
791:
792: \bibitem[Buote et al.(2002)]{Buo02} Buote, D.\ A., Jeltema, T.\ E., Canizares, C.\ R.,
793: \& Garmire, G.\ P.\ 2002, ApJ, 577, 183
794:
795: \bibitem[Diemand et al.(2007)]{Diemand07} Diemand, J., Kuhlen, M., \& Madau, P.\ 2007,
796: ApJ, 657, 262
797:
798: \bibitem[Dubinski(1994)]{Dub94} Dubinski, J.\ 1994, ApJ, 431, 617
799:
800: \bibitem[Edelmann et al.(2005)]{Edelmann05} Edelmann, H., Napiwotzki, R., Heber, U., Christlieb, N., \& Reimers, D.\ 2005, ApJ, 634, L181
801:
802: \bibitem[Eisenhauer et al.(2005)]{Eisenhauer05} Eisenhauer, F., et al.\ 2005, ApJ, 628, 246
803:
804: \bibitem[Ghez et al.(2005)]{Ghez05} Ghez, A.\ M., et al.\ 2005, ApJ, 620, 744
805:
806: \bibitem[Ginsburg \& Loeb(2006)]{GL06} Ginsburg, I., \& Loeb, A.\ 2006, MNRAS, 368,221
807:
808: \bibitem[Gnedin et al.(2005)]{Gnedin05} Gnedin, O.\ Y., Gould, A., Miralda-Escud\'{e}, J., \& Zentner, A.\ R.\ 2005, ApJ, 634, 344
809:
810: \bibitem[Gualandris \& Portegies Zwart(2007)]{GPZ07} Gualandris, A., \& Portegies Zwart, S.\ 2007, MNRAS, in press (astro-ph/0612673)
811:
812: \bibitem[Hayashi et al.(2006)]{HNS06} Hayashi, E., Navarro, J.\ F., \& Springel, V.\ 2006,
813: MNRAS, in press (astro-ph/0612327)
814:
815: \bibitem[Helmi(2004)]{Helmi04} Helmi, A.\ 2004, MNRAS, 351, 643
816:
817: \bibitem[Hernquist(1990)]{Hern90} Hernquist, L.\ 1990, ApJ, 356, 359
818:
819: \bibitem[Hills(1988)]{H88} Hills, J.\ G.\ 1988, Nature, 331, 687
820:
821: \bibitem[Hirsch et al.(2005)]{Hir05} Hirsch, H.\ A., Heber, U., O'Toole, S.\ J., \&
822: Bresolin, F.\ 2005, A\&A, 444, L61
823:
824: \bibitem[Hoekstra et al.(2004)]{Hoe04} Hoekstra, H., Yee, H.\ K.\ C., \& Gladders,
825: M.\ D.\ 2004, ApJ, 606, 67
826:
827: \bibitem[Ibata et al.(2001)]{Ibata01} Ibata, R., Lewis, G.\ F., Irwin, M., Totten, E.,
828: \& Quinn, T.\ 2001, ApJ, 551, 294
829:
830: \bibitem[Jing \& Suto(2002)]{JS02} Jing, Y.\ P., \& Suto, Y.\ 2002, ApJ, 574, 538
831:
832: \bibitem[Kazantzidis et al.(2004)]{Kazant04} Kazantzidis, S., Kravtsov, A.\ V., Zentner,
833: A.\ R., Allgood, B., Nagai, D., \& Moore, B.\ 2004, ApJ, 611, L73
834:
835: \bibitem[Klypin et al.(1999)]{Klypin99} Klypin, A.\ A., Kravtsov, A.\ V., Valenzuela, O., \&
836: Prada, F.\ 1999, ApJ, 630, L141
837:
838: \bibitem[Levin(2006)]{Levin06} Levin, Y.\ 2006, ApJ, 653, 1203
839:
840: \bibitem[Miyamoto \& Nagai(1975)]{MN75} Miyamoto, M., \& Nagai, R.\ 1975, PASJ, 27, 533
841:
842: \bibitem[Moore et al.(1999)]{Moore99} Moore, B., Ghigna, S., Governato, F.,
843: Lake, G., Quinn, T., Stadel, J., \& Tozzi, P.\ 1999, ApJ, 524, L19
844:
845: \bibitem[Moore et al.(2004)]{Moore04} Moore, B., Kazantzidis, S., Diemand, J.,
846: \& Stadel, J.\ 2004, MNRAS, 354, 522
847:
848: \bibitem[Navarro et al.(1996)]{NFW96} Navarro, J.\ F., Frenk, C.\ S., \& White, S.\ D.\ M.,
849: 1996, ApJ, 462, 563
850:
851: \bibitem[O'Leary \& Loeb(2006)]{OL06} O'Leary, R.\ M., \& Loeb, A.\ 2006,
852: MNRAS, submitted (astro-ph/0609046)
853:
854: \bibitem[Quinlan(1996)]{Q96} Quinlan, G.\ D.\ 1996, New Astron., 1, 35
855:
856: \bibitem[Sesana et al.(2006)]{SHM06} Sesana, A., Haardt, F., \& Madau, P.\ 2006,
857: ApJ, 651, 392
858:
859: \bibitem[Sesana et al.(2007)]{SHM07} Sesana, A., Haardt, F., \& Madau, P.\ 2007,
860: ApJ, in press (astro-ph/0612265)
861:
862: \bibitem[Smith et al.(2006)]{Smith06} Smith, M.\ C., et al.\ 2006, MNRAS,
863: submitted (astro-ph/0611671)
864:
865: \bibitem[Yu(2002)]{Y02} Yu, Q.\ 2002, MNRAS, 331, 935
866:
867: \bibitem[Yu \& Tremaine(2003)]{YT03} Yu, Q., \& Tremaine, S.\ 2003, ApJ, 599, 1129
868:
869: \bibitem[Zier \& Biermann(2001)]{ZB01} Zier, C., \& Biermann, P.\ L.\ 2001,
870: A\&A, 377, 23
871:
872: \end{thebibliography}
873:
874: \end{document}
875: