1: \documentclass[onecolumn]{emulateapj}
2:
3: \bibliographystyle{apj}
4: \submitted{ApJ, in press}
5:
6: \shorttitle{Massive disks in triaxial halos}
7: \shortauthors{Bailin et~al.}
8:
9:
10: \defcitealias{jog00}{J2K}
11: \defcitealias{hn06}{HN06}
12: \defcitealias{hns07}{HNS}
13:
14: \newcommand{\Msun}{{\ensuremath{\mathrm{M}_{\sun}}}}
15: \newcommand{\fpert}{\ensuremath{f_{\mathrm{pert}}}}
16: \newcommand{\epsdisk}{{\ensuremath{\epsilon_{\mathrm{disk}}}}}
17: \newcommand{\diffd}{\ensuremath{\mathrm{d}}}
18: \newcommand{\vesc}{\ensuremath{V_{\mathrm{esc}}}}
19:
20:
21: \begin{document}
22:
23: \title{Self-consistent massive disks in
24: triaxial dark matter halos}
25:
26: \author{Jeremy Bailin\altaffilmark{1,2}, Joshua D. Simon\altaffilmark{3},
27: Alberto D. Bolatto\altaffilmark{4}, Brad K. Gibson\altaffilmark{5}
28: and Chris Power\altaffilmark{1}}
29: \email{jbailin@astro.swin.edu.au}
30: \altaffiltext{1}{Centre for Astrophysics and Supercomputing,
31: Swinburne University of Technology, Mail H39, PO Box 218, Hawthorn, Victoria,
32: 3122, Australia}
33: \altaffiltext{2}{Current address: Department of Physics \&\ Astronomy, ABB-241,
34: McMaster University, 1280 Main St. W, Hamilton, ON L8S 4M1, Canada}
35: \altaffiltext{3}{Department of Astronomy, California Institute of Technology,
36: 1200 East California Boulevard, MS 105-24, Pasadena, CA 91125, USA}
37: \altaffiltext{4}{Department of Astronomy and Radio Astronomy Laboratory,
38: University of California at Berkeley, Berkeley, CA 94720, USA}
39: \altaffiltext{5}{Centre for Astrophysics, University of Central Lancashire,
40: Preston, PR1 2HE, UK}
41:
42:
43: \begin{abstract}
44: Galactic disks in triaxial dark matter halos become deformed
45: by the elliptical potential in the plane of the disk in such a way as to
46: counteract the halo ellipticity.
47: We develop a technique to calculate the equilibrium configuration
48: of such a disk in the combined disk-halo potential, which is
49: based on the method of \citet{jog00} but accounts
50: for the radial variation in both the halo potential
51: and the disk ellipticity.
52: This crucial ingredient results in qualitatively
53: different behavior of the disk: the disk circularizes
54: the potential at small radii, even for a
55: reasonably low disk mass. This effect has important implications
56: for proposals to reconcile cuspy
57: halo density profiles with low surface brightness galaxy
58: rotation curves using halo triaxiality.
59: The disk ellipticities in our models are consistent with
60: observational estimates based on two-dimensional
61: velocity fields and isophotal axis ratios.
62: \end{abstract}
63:
64: \keywords{%
65: galaxies: kinematics and dynamics ---
66: galaxies: halos ---
67: galaxies: spiral ---
68: galaxies: structure ---
69: methods: numerical ---
70: dark matter}
71:
72: \section{Introduction}\label{intro section}
73: Galaxies are thought to be surrounded by large dark matter halos.
74: These halos are much more massive than the visible components
75: of galaxies, and dominate much of the dynamics.
76: Although dark matter halos are often assumed to be spherical
77: for simplicity,
78: the halos that form in
79: cosmological simulations
80: are quite flattened, with typical intermediate axis ratios
81: of $b/a \sim 0.8$ and minor axis ratios of $c/a \sim 0.6$,
82: with some systematic
83: variation depending on the mass of the halo and the
84: radius at which the shape is measured
85: \citep{warren-etal92,js02,bs05-alignment,allgood-etal06}.
86: This non-sphericity is a testable prediction of
87: cosmological models.
88:
89: Simulations of disk galaxy formation within dark matter halos
90: find that the presence of the disk modifies the shape
91: of the halo, reducing the halo triaxiality
92: \citep{dubinski94,kazantzidis-etal04,bailin-etal05-diskhalo,bs06}.
93: However,
94: as long as the final shape of the halo retains some ellipticity in the
95: plane of the disk, the dynamics
96: and shape of the disk will be affected by
97: the deviations from axisymmetry \citep[e.g.][]{gv86,sfdz97}.
98:
99: Observations indicate that many disks
100: do indeed have small but non-zero
101: ellipticities. Evidence for elliptical disks
102: comes from harmonic decomposition of galaxy photometry
103: \citep{rz95}, harmonic decomposition
104: of two-dimensional velocity fields \citep{sfdz97,simon-etal05},
105: and statistical analysis of the distribution of projected
106: shapes \citep{ryden06}.
107: These results qualitatively confirm that galactic dark matter
108: halos are elliptical; precision measurements could provide
109: direct constraints on the shapes of the halos.
110:
111: Recently, \citet[][hereafter HN06]{hn06} proposed that elliptical orbits within
112: the disk produced by the triaxiality of the halo could reconcile
113: cuspy density profiles that form in cosmological simulations
114: \citep[][hereafter NFW]{nfw96} with observed rotation curves
115: of low surface brightness (LSB) galaxies, which often
116: appear to require halos with constant density cores
117: \citep[e.g.,][]{deblok-etal01}.
118: This analysis did not take into account the self-gravity of the disk. In
119: galaxies with massive disks, the gravity of the disk contributes to
120: the net potential and the dynamics of the disk are determined by a
121: combination of the halo and the disk itself. In order to draw
122: conclusions about the shape of the halo from the measured dynamics of
123: the disk, we must determine self-consistently both how the disk is
124: perturbed by the potential and how the perturbed disk contributes to
125: the potential.
126:
127: An elegant method to carry out these calculations was proposed by \citet{jog00} (hereafter
128: \citetalias{jog00}; see also \citealt{jog97,jog99}). By assuming a
129: logarithmic halo potential with a small constant elliptical perturbation
130: and an exponential disk with a small constant elliptical response,
131: \citetalias{jog00} solved for the self-consistent response. She
132: demonstrated that the disk response dilutes the ellipticity
133: of the potential most strongly at $1.42$ disk scale lengths.
134:
135: There are a number of simplifying assumptions in \citetalias{jog00} that
136: require examination. The most important
137: assumption is that both the halo perturbation and the disk response
138: are constant with radius.
139: In contrast, \citetalias{hn06} demonstrated
140: that a radially-varying perturbation is required to reconcile LSB long-slit
141: rotation curves with cuspy halo profiles.
142: Indeed, cosmological simulations predict a radially-varying
143: perturbation in the halo potential;
144: even if halos had isodensity surfaces of constant ellipticity,
145: the shape of the potential would vary with radius,
146: and \citet[][hereafter HNS]{hns07},
147: who directly measured the shapes of isopotential surfaces
148: of cosmological halos,
149: found even stronger variation with radius.
150: The response of the disk is also not expected to be uniform
151: in a radially-varying potential.
152:
153:
154: In this paper, we generalize the method of \citetalias{jog00} to the
155: more realistic case of radially varying halo perturbations and radially
156: varying disk responses. In \S~\ref{method section} we detail the method
157: for determining the disk shape and dynamics. In \S~\ref{results section}
158: we use this method to determine the shapes and dynamics of disks in
159: sample triaxial halos and demonstrate how the results depend
160: on the properties of the disks and halos.
161: \S~\ref{discussion section} discusses our results in the
162: context of observations that directly probe disk ellipticity, and in
163: \S~\ref{conclusions section} we present our conclusions.
164:
165:
166: \section{Method}\label{method section}
167:
168: \subsection{Outline}\label{outline method section}
169:
170: Our method, which is based closely on \citetalias{jog00},
171: is as follows:
172: \begin{enumerate}
173: \item Calculate the axisymmetric component of the potential and the
174: elliptical perturbation in the potential induced by the triaxial halo
175: (\S~\ref{axisymmetric potential section}).
176: \item Calculate the closed orbits and corresponding disk ellipticity
177: for a given net perturbation to the potential
178: (\S\S~\ref{closed orbit section} and \ref{disk ellipticity section}).
179: \item Calculate the elliptical perturbation in the potential induced by
180: a given disk ellipticity (\S~\ref{disk potential section}).
181: \item Solve for the form of the net potential perturbation that satisfies
182: all of the above constraints (\S~\ref{self-con soln section}).
183: \end{enumerate}
184: Throughout this procedure, the halo is kept fixed, i.e. it does not
185: respond to the presence of the disk. Therefore, the potential that should be
186: used in the calculation is the real shape of the halo in which
187: the disk lies, which is less triaxial than the shape the halo would
188: have in the absence of baryonic processes
189: \citep{kazantzidis-etal04,bailin-etal05-diskhalo,bs06}.
190:
191: The main differences between this work and \citetalias{jog00} are:
192: \begin{enumerate}
193: \item Where \citetalias{jog00} assumes a logarithmic potential for the halo,
194: we evaluate the radial form of the potential directly from
195: a density distribution motivated by cosmological simulations.
196: \item Where \citetalias{jog00} assumes a constant perturbation
197: to the halo potential, we allow the perturbation to vary with
198: radius and either evaluate it directly from a triaxial density
199: distribution motivated by cosmological simulations or
200: use parametrizations developed from measurements of halos
201: in cosmological simulations.
202: \item Where \citetalias{jog00} assumes that the disk responds with a constant
203: ellipticity, we allow the disk ellipticity to vary with radius.
204: \end{enumerate}
205: Whenever we carry out numerical calculations in this paper,
206: all radially-varying functions are tabulated on a radial grid
207: sampled at $100$ radii $R_i$ spaced logarithmically between $0.1$
208: and $100$~kpc in order to finely sample the inner region of the disk
209: where the quantities vary most rapidly; $50$ grid points lie at
210: $R<1$~kpc. The functions are linearly interpolated
211: between grid points when their values are required at arbitrary
212: radii.
213:
214:
215: \subsection{Axisymmetric potential and halo perturbation}%
216: \label{axisymmetric potential section}
217: For the perturbative approach, we assume that within the plane of the disk
218: (which we take to be $z=0$ for simplicity), the total potential can be written as
219: \begin{eqnarray}
220: \Phi(R,\theta) &=& \Phi_0(R) + \Phi_{\mathrm{pert}}(R,\theta)\nonumber\\
221: &=& \Phi_0(R) \left(1 + \fpert(R) \cos m\theta\right),
222: \label{eq fpert defn}
223: \end{eqnarray}
224: where $R$, $\theta$, and $z$ are the cylindrical coordinates.
225: For the purposes of this paper, we will assume an elliptical
226: perturbation (i.e., $m=2$) from now on
227: (see Appendix~\ref{second order terms section} for a discussion of
228: the $m=4$ mode).
229: We assume that $\fpert(R)$ is small and varies slowly with $R$.
230: Note that for $\fpert > 0$, isopotentials are elongated along the $x$ axis
231: and closed orbits in the disk are elongated along the $y$ axis.
232:
233: Both the disk and halo contribute to both the axisymmetric and $m=2$ components
234: of the potential:
235: \begin{equation}
236: \Phi_0(R) = \Phi_0^{\mathrm{halo}}(R) + \Phi_0^{\mathrm{disk}}(R),
237: \end{equation}
238: \begin{equation}
239: \fpert(R) = \fpert^{\mathrm{halo}}(R) + \fpert^{\mathrm{disk}}(R).
240: \label{eq fpert breakdown}
241: \end{equation}
242:
243: To first order,
244: the $m=2$ perturbation in the potential induces
245: an $m=2$ perturbation in the surface density distribution of the
246: otherwise exponential disk
247: (see Appendix~\ref{second order terms section} for a justification of
248: our decision to neglect the higher-order terms):
249: \begin{equation}
250: \Sigma(R,\theta) = \Sigma_0 \exp \left[ - \frac{R}{R_d}
251: \left( 1 - \frac{\epsdisk(R)}{2} \cos 2\theta \right) \right]
252: \label{disk density distribution}
253: \end{equation}
254: We assume $\epsdisk(R)$, the ellipticity of the isodensity ellipse,
255: is small and varies slowly with $R$.
256: The axisymmetric component of the disk potential is given by
257: \begin{equation}
258: \Phi_0^{\mathrm{disk}}(R) = - \pi G \Sigma_0 R \left[ I_0(y)
259: K_1(y) - I_1(y) K_0(y) \right]
260: \end{equation}
261: ($y=R/2 R_d$; see \citealt{freeman70}; \citealt{BT}, eq. 2-168), where
262: $I$ and $K$ are modified Bessel functions.
263:
264: Given the density distribution of the dark matter halo,
265: both axisymmetric and $m=2$ components of the halo potential can be
266: evaluated.
267: If the isodensity surfaces of the halo are self-similar ellipsoids,
268: then the halo density can be written as:
269: \begin{equation}
270: \rho(x,y,z) = \rho(s),
271: \end{equation}
272: where
273: \begin{equation}
274: s^2 = \left(\frac{x}{a}\right)^2 + \left(\frac{y}{b}\right)^2 +
275: \left(\frac{z}{c}\right)^2.
276: \end{equation}
277: For example, we can use an NFW form for the density:
278: \begin{equation}
279: \rho(s) = \frac{\rho_0}{\frac{s}{r_s}\left(1 + \frac{s}{r_s}\right)^2}
280: \label{NFW density eq}
281: \end{equation}
282: \citep{nfw96,js02}.
283: We make no assumptions about the relative magnitudes
284: of $a$, $b$, and $c$; they are simply the relative axis ratios along the $x$,
285: $y$, and $z$ axes respectively. Therefore, the disk,
286: which lies in the $xy$ plane, can be oriented in
287: any of the principal planes of the halo.
288:
289: We calculate the halo potential along the $x$ axis,
290: $\Phi_x^{\mathrm{halo}}(R) \equiv \Phi^{\mathrm{halo}}(R,0,0)$,
291: and along the $y$ axis,
292: $\Phi_y^{\mathrm{halo}}(R) \equiv \Phi^{\mathrm{halo}}(0,R,0)$,
293: by numerically integrating
294: eq. (2-99) of \citet{BT} using $\rho(s)$
295: given in equation (\ref{NFW density eq}).
296: This allows us to calculate $\Phi_0^{\mathrm{halo}}$ and
297: $\fpert^{\mathrm{halo}}$ as
298: \begin{equation}
299: \Phi_0^{\mathrm{halo}}(R) = \frac{1}{2} \left[\Phi_x(R) + \Phi_y(R)\right],
300: \end{equation}
301: and
302: \begin{equation}
303: \fpert^{\mathrm{halo}}(R) =
304: \frac{1}{2 \Phi_0(R)} \left[\Phi_x(R) - \Phi_y(R)\right].
305: \label{eq fperthalo}
306: \end{equation}
307:
308: \subsection{Closed orbits}\label{closed orbit section}
309: In a triaxial potential,
310: dissipative gas settles on stable closed loop orbits when
311: such orbits exist \citep{elzant-01}. This is the case throughout
312: the potential of a centrally-concentrated mass profile such
313: as the NFW profile. Therefore, the structure of a galactic
314: disk, which consists of gas clouds and stars formed within those gas clouds,
315: is determined by the form of the closed orbits.
316: These have been examined
317: in detail by \citet{sfdz97}, and simplified into a convenient
318: form by \citetalias{hn06}.
319: These previous derivations have assumed that the perturbation to the
320: potential is constant over the radial excursion of an orbit, which is
321: not the case for the radially-varying perturbations that we wish to
322: study. We have therefore rederived the equations for closed orbits
323: within a radially-varying perturbation from the equations of motion.
324: The orbits follow
325: \begin{equation}
326: R = R_0 \left( 1 - \frac{\fpert\, a_{12}}{2}
327: \cos 2 \theta_0 \right)
328: \label{closed orbit R}
329: \end{equation}
330: \begin{equation}
331: \theta = \theta_0 + \frac{a_{12}+a_{32}}{2m}
332: \fpert \sin 2 \theta_0,
333: \label{closed orbit theta}
334: \end{equation}
335: with velocities
336: \begin{equation}
337: V_R = V_c\, \fpert\, a_{12} \sin 2 \theta_0 \label{vr defn}
338: \label{closed orbit VR}
339: \end{equation}
340: \begin{equation}
341: V_{\theta} = V_c \left( 1 + \frac{\fpert\, a_{32}}{2} \cos 2 \theta_0
342: \right), \label{vtheta defn}
343: \label{closed orbit Vtheta}
344: \end{equation}
345: where $R_0$ and $\theta_0$ define the guiding center of the orbit,
346: $\theta_0=\Omega_0\, t$, and the following
347: functions of $\Phi_0(R)$ are evaluated at $R_0$:
348: \begin{equation}
349: \Omega_0(R) = \sqrt{\frac{1}{R} \frac{\diffd \Phi_0}{\diffd R}}
350: \end{equation}
351: \begin{equation}
352: V_c(R) = R\, \Omega_0(R)
353: \end{equation}
354: \begin{equation}
355: \vesc(R) = \sqrt{2 |\Phi_0|}
356: \end{equation}
357: \begin{equation}
358: g_m(R) = \frac{1}{\Omega_0^2} \frac{\diffd^2 \Phi_0}{\diffd R^2} - (m^2 - 3)
359: \end{equation}
360: \begin{equation}
361: a_{1m}(R) = \frac{2}{g_m(R)}\left[ 1 - \frac{\vesc^2}{V_c^2}
362: \left(1 + \frac{1}{2} \frac{R}{\fpert} \frac{\diffd \fpert}{\diffd R}\right)
363: \right]
364: \label{a1m definition}
365: \end{equation}
366: \begin{equation}
367: a_{3m}(R) = a_{1m}(R) + \frac{\vesc^2}{V_c^2}.
368: \end{equation}
369: The coefficients $a_{1m}$ and $a_{3m}$ quantify the degree
370: to which the radius $R$ and angular velocity $V_{\theta}$ respectively,
371: which are constant for a circular orbit, vary for a unit
372: perturbation to the potential.
373: Our expressions differ from those in \citetalias{hn06} for the following
374: reasons:
375: (1) we have taken the radial variation of the perturbation into account
376: in our derivation of equation (\ref{a1m definition});
377: (2) we have generalized the expression for $g_m(R)$ to be valid for
378: all $m$, while the expression in \citetalias{hn06} is specific to $m=2$;
379: (3) we have removed the factor of \fpert\ from the definitions of $a_{1m}$
380: and $a_{3m}$ for convenience later (note that these quantities still
381: depend implicitly on \fpert\ through its derivative);
382: and (4) we have generalized the expression for $a_{3m}$ to be valid
383: for all potential profiles, while the expression in \citetalias{hn06} is
384: only valid when $g_2(R)=-3$, which is not the case in the inner
385: regions of an NFW potential.
386:
387: Given the tabulated values of $\Phi_0$, these functions can be evaluated
388: at the grid points $R_i$. Because $\Phi_0$ has been calculated from
389: analytic functions, the tabulated values are relatively free of
390: noise and even the numerical second derivative does not contain large
391: fluctuations.
392: Note that the closed orbits are elongated \emph{perpendicular} to the
393: isopotential contours.
394:
395:
396: \subsection{Disk ellipticity}\label{disk ellipticity section}
397: The disk must satisfy the continuity equation. In cylindrical
398: coordinates:
399: \begin{equation}
400: \frac{\partial}{\partial R} \left[ R\, \Sigma(R,\theta)\,
401: V_R(R,\theta) \right] +
402: \frac{\partial}{\partial\theta} \left[ \Sigma(R,\theta)\,
403: V_{\theta}(R,\theta) \right] = 0.
404: \end{equation}
405: We substitute $\Sigma(R,\theta)$ from (\ref{disk density distribution}),
406: $V_R(R,\theta)$ from (\ref{vr defn}), and $V_{\theta}(R,\theta)$ from
407: (\ref{vtheta defn}). To first order in the small quantities
408: \fpert, \epsdisk, and their derivatives:
409: \begin{equation}
410: \frac{R}{R_d} \epsdisk(R) = \fpert(R) \left[ a_{12}(R) \left(1 -
411: \frac{R}{R_d} + \frac{R}{V_c} \frac{\diffd V_c}{\diffd R} \right)
412: - a_{32}(R) \right].
413: \label{epsdisk fpert a12 reln}
414: \end{equation}
415: The neglected second-order terms induce
416: small $m=4$ perturbations in the disk;
417: see Appendix~\ref{second order terms section} for details.%
418: \footnote{We have also omitted the term
419: in equation (\ref{epsdisk fpert a12 reln}) that is proportional
420: to $R\> \diffd(\fpert\, a_{12})/\diffd R$; however, its
421: effect is negligible.}
422: Equation~(\ref{epsdisk fpert a12 reln})
423: provides us with a relationship between the radial profile of
424: the potential (embodied in $a_{12}$, $a_{32}$ and $V_c$), the strength of
425: the perturbation in the net potential (\fpert), and
426: the ellipticity of the disk (\epsdisk).
427:
428:
429: \subsection{Disk perturbation potential}\label{disk potential section}
430: The ellipticity of the disk generates an $m=2$ perturbation to the
431: disk potential.
432: We calculate this as follows:
433: \begin{eqnarray}
434: \Phi^{\mathrm{disk}}_{\mathrm{pert}} \equiv
435: \Phi^{\mathrm{disk}} - \Phi_0^{\mathrm{disk}} =
436: -G \sum_{m=-\infty}^{\infty} \exp(i m \theta)
437: \int_0^\infty J_m(k R)\, \exp(-k\, |z|)\, \diffd k
438: \nonumber\\
439: \int_0^\infty J_m(k R')\, R'\, \diffd R'
440: \int_0^{2\pi} \left[\Sigma(R',\theta') - \Sigma_0\, \exp(-R'/R_d)\right]
441: \, \exp(-i m \theta')\, \diffd \theta'
442: \label{full disk integral}
443: \end{eqnarray}
444: (\citealp{BT}, eq. 2P-8).
445: We restrict ourselves to the plane $z=0$, corresponding to an
446: infinitely thin disk.
447:
448: For small perturbations, the perturbed surface density is
449: \begin{equation}
450: \Sigma(R',\theta') - \Sigma_0\, e^{-R'/R_d} \approx
451: \Sigma_0\, e^{-R'/R_d}
452: \frac{R'}{R_d} \frac{\epsdisk(R')}{2} \cos 2\theta'.
453: \label{sigma pert}
454: \end{equation}
455: Substituting (\ref{sigma pert}) into (\ref{full disk integral}),
456: we note that the integral over $\diffd \theta'$ vanishes except
457: when $m=\pm 2$.
458: Since $\int_0^{2\pi} \cos 2\theta'\, e^{\pm i 2 \theta'}\, \diffd\theta'
459: =\pi$, $e^{ix}+e^{-ix}=2\cos x$, and $J_2(x) = J_{-2}(x)$, we find
460: \begin{equation}
461: \Phi^{\mathrm{disk}}_{\mathrm{pert}}(R,\theta)
462: = -\pi G\, \Sigma_0 \cos 2 \theta
463: \int_0^{\infty} J_2(k R)\,\diffd k
464: \int_0^{\infty} J_2(k R')
465: R'\, \exp\left(-\frac{R'}{R_d}\right)
466: \frac{R'}{R_d} \epsdisk(R')\, \diffd R'.
467: \end{equation}
468: We can express $(R'/R_d)\, \epsdisk(R')$ in the final integral in terms
469: of \fpert, $a_{12}$, $a_{32}$, and $V_c$ using (\ref{epsdisk fpert a12 reln}):
470: \begin{eqnarray}
471: \Phi^{\mathrm{disk}}_{\mathrm{pert}}(R,\theta)
472: = -\pi G\, \Sigma_0 \cos 2 \theta
473: \int_0^{\infty} J_2(k R)\,\diffd k
474: \int_0^{\infty} J_2(k R')\, R'\, \exp\left(-\frac{R'}{R_d}\right)
475: \nonumber\\
476: \fpert(R')
477: \left[a_{12}(R') \left( 1 - \frac{R'}{R_d} +
478: \frac{R'}{V_c(R')}\frac{\diffd V_c(R')}{\diffd R'}\right) -
479: a_{32}(R')\right]
480: \, \diffd R'.
481: \end{eqnarray}
482:
483: Because we do not know \textit{a priori} the net perturbation
484: $\fpert(R)$, we cannot immediately evaluate these integrals.
485: However, if we can find a function
486: $\fpert^{\mathrm{proxy}}$ whose form is similar to $\fpert$,
487: i.e. if $\fpert/\fpert^{\mathrm{proxy}}$
488: is a slowly-varying function of $R$, then we can approximate
489: the potential as:
490: \begin{eqnarray}
491: \Phi^{\mathrm{disk}}_{\mathrm{pert}}(R,\theta) \approx
492: -\frac{\fpert(R)}{\fpert^{\mathrm{proxy}}(R)} \pi G\, \Sigma_0 \cos 2 \theta
493: \int_0^{\infty} J_2(k R)\,\diffd k
494: \int_0^{\infty} J_2(k R')\, R'\, \exp\left(-\frac{R'}{R_d}\right)
495: \nonumber\\
496: \fpert^{\mathrm{proxy}}(R')
497: \left[a_{12}(R') \left( 1 - \frac{R'}{R_d} +
498: \frac{R'}{V_c(R')}\frac{\diffd V_c(R')}{\diffd R'}\right) -
499: a_{32}(R')\right]
500: \, \diffd R'.
501: \label{eq fproxy integral}
502: \end{eqnarray}
503: Note that $a_{12}(R')$ and $a_{32}(R')$ depend implicitly on
504: $\fpert^{\mathrm{proxy}}(R')$ through its derivative.
505: A first approximation can be obtained by setting
506: $\fpert^{\mathrm{proxy}} = \fpert^{\mathrm{halo}}$, whose values
507: have been tabulated from (\ref{eq fperthalo}).
508: Because the integral over $R'$ is independent of $R$
509: and the integral over $k$ is independent of $R'$,
510: the integrals can be evaluated independently on fine grids
511: of $k$ and $R$, respectively.
512: Using this technique, we calculate
513: \begin{eqnarray}
514: \eta(R) \equiv \frac{\pi G\, \Sigma_0}{\fpert^{\mathrm{proxy}}(R)}
515: \int_0^{\infty} J_2(k R)\, \diffd k
516: \int_0^{\infty} J_2(k R')\, R'\, \exp\left(-\frac{R'}{R_d}\right)
517: \nonumber\\
518: \fpert^{\mathrm{proxy}}(R')
519: \left[a_{12}(R') \left( 1 - \frac{R'}{R_d} +
520: \frac{R'}{V_c(R')}\frac{\diffd V_c(R')}{\diffd R'}\right) -
521: a_{32}(R')\right]
522: \, \diffd R'
523: \label{eq etar defn}
524: \end{eqnarray}
525: at each grid point $R_i$. The perturbation potential due to the disk
526: is then given by
527: \begin{equation}
528: \Phi^{\mathrm{disk}}_{\mathrm{pert}}(R,\theta) = -\fpert(R)\,
529: \eta(R)\, \cos 2\theta.
530: \label{eq phi disk resp}
531: \end{equation}
532: Expressed in this form, the meaning of $\eta(R)$ becomes clear: it
533: is the magnitude of the disk response to a unit perturbation
534: in the potential.
535:
536:
537: \subsection{Self-consistent solution}\label{self-con soln section}
538: For clarity, we repeat here the important equations:
539: \begin{eqnarray}
540: \label{eq phipert vs phi0}
541: \Phi_{\mathrm{pert}}(R,\theta) &=& \Phi_0(R)\, \fpert(R)\, \cos 2\theta \\
542: &=&
543: \Phi_{\mathrm{pert}}^{\mathrm{halo}}(R,\theta) +
544: \Phi_{\mathrm{pert}}^{\mathrm{disk}}(R,\theta),
545: \end{eqnarray}
546: \begin{equation}
547: \Phi_{\mathrm{pert}}^{\mathrm{halo}}(R,\theta) = \fpert^{\mathrm{halo}}(R)\,
548: \Phi_0(R)\, \cos 2\theta,
549: \end{equation}
550: (see eqs. \ref{eq fpert defn} and \ref{eq fpert breakdown}), and
551: \begin{equation}
552: \Phi_{\mathrm{pert}}^{\mathrm{disk}}(R,\theta) = -\fpert(R)\,
553: \eta(R)\, \cos 2\theta,
554: \label{eq fpert repeat}
555: \end{equation}
556: (repeated from eq. \ref{eq phi disk resp}).
557: The physical interpretation of these equations is that
558: the disk response is proportional to the net perturbation \fpert,
559: which is itself the sum of the disk response and the imposed halo
560: perturbation. The disk response is opposite in sign to the halo
561: perturbation,
562: so \fpert\ in the self-consistent solution must
563: be reduced with respect to the imposed halo perturbation.
564:
565: The self-consistent solution can be obtained by collection
566: equations (\ref{eq phipert vs phi0}) through (\ref{eq fpert repeat}):
567: \begin{equation}
568: \fpert(R) = \fpert^{\mathrm{halo}}(R) \frac{1}{1 + \eta(R)/\Phi_0(R)}.
569: \end{equation}
570: In other words, the response of the disk causes the overall potential
571: perturbation to be reduced by a factor
572: of $1 + \eta(R)/\Phi_0(R)$. All terms on the right hand side have
573: been tabulated at the grid points $R_i$, resulting in a trivial
574: evaluation of $\fpert(R)$.
575:
576: Armed with this new estimate of \fpert, we can reexamine equation
577: (\ref{eq etar defn}), substitute $\fpert^{\mathrm{proxy}}=\fpert$,
578: and calculate a new value of $\eta(R)$ and therefore of $\fpert(R)$.
579: We repeat this procedure until the maximum change
580: between iterations in the quantity $\fpert/\fpert^{\mathrm{halo}}$,
581: which is a robust indicator of the relative error in \fpert,
582: is less than $10^{-3}$ at all radii; this is typically achieved
583: in $20$--$30$ iterations. We have confirmed for some specific cases
584: that our solution agrees
585: to within $\sim 2\%$ of the true solution (assumed to have converged after
586: a very large number of iterations) at all radii, and to much higher precision
587: at most radii
588: (see Figure~\ref{convergence figure}).
589: Adopting a stricter convergence criterion has no effect on our results.
590: The relatively large number of iterations is required in order to
591: accurately capture the sharp feature where $\fpert \rightarrow 0$
592: that is seen in the solutions (see \S~\ref{results section}).
593:
594: \begin{figure}
595: \plotone{f1.eps}
596: \caption{\label{convergence figure}%
597: Maximum change in the solution for $\fpert/\fpert^{\mathrm{halo}}$
598: per iteration (dashed line) and maximum
599: difference between the solution at a given iteration and the true
600: solution, assumed to have converged after $100$ iterations (solid line),
601: for the fiducial halo and $3\times 10^9~\Msun$ disk
602: of \S~\ref{disk mass results section}.
603: The vertical dotted line indicates the point
604: where our convergence criterion is achieved.}
605: \end{figure}
606:
607:
608: Given \fpert, the disk ellipticity \epsdisk\ can be calculated as a function
609: of radius directly from equation (\ref{epsdisk fpert a12 reln}) and the
610: forms of the closed orbits can be calculated from equations
611: (\ref{closed orbit R}) -- (\ref{closed orbit Vtheta}). This provides a
612: complete description of the disk.
613:
614:
615: \section{Results}\label{results section}
616: In this section, we demonstrate how the disk dilutes
617: the elliptical potential of the halo,
618: and give examples of the net ellipticity
619: induced in the disk.
620: In \S~\ref{disk mass results section}, we demonstrate the main
621: features of the disk-halo systems using a halo with constant
622: axis ratios,
623: while in \S~\ref{varying halo parameters section} we investigate
624: how these results are affected by varying disk and halo
625: parameters such as the halo concentration, axis ratio, disk
626: scale length, and run of halo axis ratio with radius.
627:
628:
629: \subsection{Response for various disk masses}
630: \label{disk mass results section}
631: We demonstrate the main features of our models
632: using a fiducial triaxial
633: NFW halo with mass $M_{200}=10^{12}~\Msun$, axis ratios
634: $b/a=0.8$ and $c/a=0.6$, and a concentration $c_{200}=12$.%
635: \footnote{$M_{200}$ and $c_{200}$ refer to the mass and
636: concentration relative to the radius $r_{200}$, defined such that
637: the mean density within $r_{200}$ is $200$ times the critical
638: density. We assume the Hubble parameter $h=0.7$.}
639: These values are typical for galaxy-sized dark matter halos
640: in cosmological simulations \citep{allgood-etal06}.
641: The disk rotation axis is aligned with the minor axis of the halo,
642: in agreement with the orientation of the halo angular momentum
643: in simulations \citep{bs05-alignment}.
644: All disks have radial scalelengths $R_d=3.0~\mathrm{kpc}$, with
645: masses that range from zero up to $10^{11}~\Msun$.
646:
647: \begin{figure}
648: \plotone{f2.eps}
649: \caption{\label{phi0 figure}%
650: \textit{(a)} Axisymmetric component of the potential, $\Phi_0(R)$, for
651: massless disks (black), and disks of mass $3 \times 10^8~\Msun$ (red),
652: $3 \times 10^9~\Msun$ (blue), $3 \times 10^{10}~\Msun$ (green),
653: and $10^{11}~\Msun$ (purple) within the fiducial halo
654: of \S~\ref{disk mass results section}.
655: \textit{(b)} Circular velocity curve within the unperturbed potential.
656: The disk contribution to the rotation curve is denoted with dashed
657: lines. Colors are the same as in \textit{(a)}.}
658: \end{figure}
659:
660: \begin{figure}
661: \plotone{f3_color.eps}
662: %\plotone{f3.eps}
663: \caption{\label{phixy figure}%
664: \textit{(a)}
665: The solid curves indicate the
666: potential along the $x$ (lower) and $y$ (upper) axis
667: for the disk of mass $3\times 10^9~\Msun$ within the fiducial
668: halo of \S~\ref{disk mass results section}. The dotted line indicates
669: the axisymmetric component of the potential. The horizontal and vertical
670: lines demonstrate how the ellipticity of the isopotential surfaces
671: and the magnitude of the perturbation respectively are calculated.
672: \textit{(b)}
673: The curves indicate the magnitude of the perturbation, \fpert\ (solid), and
674: ellipticity of the potential, $\epsilon_{\Phi}$ (dashed) for the same
675: halo as in (a).}
676: \end{figure}
677:
678: \begin{figure}
679: \epsscale{0.6}
680: \plotone{f4.eps}
681: \caption{\label{disk mass figures}%
682: \textit{(a)}
683: Magnitude of the elliptical perturbation in the potential due to
684: the triaxiality of the halo ($\fpert^{\mathrm{halo}}$, solid lines),
685: and the net perturbation after
686: including the self-consistent response of the disk
687: (\fpert, dashed lines). Different disk masses are indicated by different
688: colors as in Figure 1.
689: The dotted line shows the form of \fpert\ proposed by \citetalias{hn06}
690: to produce a rotation curve mimicking a cored isothermal
691: density profile.
692: \textit{(b)}
693: Ratio by which the input perturbation $\fpert^{\mathrm{halo}}$ becomes
694: diluted due to the self-consistent response of the disk.
695: \textit{(c)}
696: Ellipticity of the disk isodensity contours ($|\epsdisk|$, solid lines;
697: note that $\epsdisk$ is negative for positive \fpert\ in the sign
698: convention we have chosen) and of orbits within the disk
699: ($\epsilon_{\mathrm{orbit}}$, dashed lines).}
700: \end{figure}
701:
702: Figure~\ref{phi0 figure} demonstrates how the axisymmetric component
703: of the potential, $\Phi_0(R)$, and the rotation
704: curve, $V_c(R)$, vary as a function of disk mass.
705: The potential is given in units of $V_{200}^2 \equiv G\, M_{200}
706: / r_{200}$.
707: For disk masses less then $3 \times 10^9~\Msun$, the halo dominates
708: the axisymmetric component of the potential at all radii, with the disk
709: becoming increasingly more important with increasing disk mass beyond
710: this.
711: The non-axisymmetric component of the potential is demonstrated in
712: Figure~\ref{phixy figure} for the $3 \times 10^9~\Msun$ disk. Although
713: the ellipticity of isopotential surfaces, $\epsilon_{\Phi}$, rises
714: to small radii, the magnitude of the potential perturbation,
715: $\fpert^{\mathrm{halo}}$, must vanish at small radii because
716: $\Phi_x(R=0) = \Phi_y(R=0)$ while $\Phi_0(R=0)$ reaches a finite
717: value.
718:
719: Figure~\ref{disk mass figures}a demonstrates both
720: the initial halo perturbation ($\fpert^{\mathrm{halo}}$, solid lines)
721: and the net perturbation after including the
722: self-consistent response of disk (\fpert, dashed lines).
723: Note that even though the halo is identical in each case,
724: $\fpert^{\mathrm{halo}}$ is lower for higher mass disks because
725: the disk contributes to $\Phi_0$.
726:
727: \citetalias{hn06} proposed that a suitable perturbation
728: \begin{equation}
729: \fpert = f_{\mathrm{iso}}(R) \approx a\, x\, e^{-x/b}
730: \end{equation}
731: (with $x \equiv R/r_s$, $a=0.1$, and $b=0.098$) would cause
732: the rotation curve of a perturbed NFW profile to mimic that of a
733: cored isothermal profile. We denote this as a dotted line in
734: Figure~\ref{disk mass figures}a.
735: The form of this perturbation is very different from the
736: form of the perturbation that we find to be induced by a triaxial halo
737: of uniform axis ratio,
738: particularly once the self-consistent response of the disk is
739: taken into account.
740:
741: Figure~\ref{disk mass figures}b demonstrates the degree to which the initial
742: halo perturbation $\fpert^{\mathrm{halo}}$ is diluted by the
743: self-consistent response of the disk. This is equal to
744: $\fpert/\fpert^{\mathrm{halo}}$ and is determined from
745: $1/(1 + \eta(R)/\Phi_0(R))$.
746: The ellipticity in the potential vanishes in the central region
747: where $\eta(R) \gg \Phi_0(R)$.
748: Even for a negligible disk mass, the potential in the
749: innermost region is circularized. This region is larger
750: for more realistic disks, which have a
751: significant impact on the potential out to several disk scale lengths.
752: As the disk mass increases, the form of the disk dominates
753: both $\eta(R)$ and $\Phi_0(R)$, and therefore this
754: function approaches an asymptotic form.
755:
756: Comparison between Figure~\ref{disk mass figures}b
757: and the equivalent figs.~2 and 3 of
758: \citetalias{jog00} reveals dramatically different behavior at small radii:
759: in \citetalias{jog00}, the ``reduction factor'' reaches a minimum at $1.42~R_d$
760: ($=4.26$~kpc for $R_d=3.0$~kpc) and then rises to unity,
761: while in Figure~\ref{disk mass figures}b
762: it falls monotonically to vanish at small radii.
763: This is a direct result of the radial variation of \fpert\
764: in a physically realistic elliptical halo.
765: Because $\eta(R)$ depends inversely on
766: \fpert\ (see eq.~\ref{eq etar defn}), which must vanish at small
767: radii, $\eta(R)$ must dominate over $\Phi_0(R)$ in the inner
768: regions and the potential must become completely circularized.
769:
770: In Figure~\ref{disk mass figures}c we plot the ellipticity of
771: the disk isodensity contours ($\epsdisk$, solid lines) and of the
772: orbits within the disk ($\epsilon_{\mathrm{orbit}} \equiv \fpert\, a_{12}$,
773: dashed lines).
774: For a massless disk, we recover the
775: results of \citetalias{hn06} that the ellipticity rises towards
776: the center of the halo.
777: However, the presence of a massive disk changes the situation
778: dramatically. Because $\fpert/\fpert^{\mathrm{halo}}$ vanishes at small
779: radii (Figure~\ref{disk mass figures}b), the equilibrium disk is
780: axisymmetric at small radii. Even very low mass disks, which contribute
781: negligibly to $\Phi_0$,
782: still respond strongly enough to the elliptical potential
783: to cause an important change in the behavior at small radii.
784: We also note that the ellipticity of disk isophotes are always
785: greater than ellipticity of orbits within the disk.
786:
787:
788: \subsection{Varying disk and halo parameters}
789: \label{varying halo parameters section}
790: \begin{figure}
791: \epsscale{0.6}
792: \plotone{f5.eps}
793: \caption{\label{hns figures}%
794: As in Figure~\ref{disk mass figures}
795: for disks in a halo potential with the same
796: radial variation of its axis ratios as halo G4 of \citetalias{hns07}.}
797: \end{figure}
798: \citetalias{hns07} found that the radial variation of the shape of
799: the potential of cosmological $N$-body halos
800: is not consistent with self-similar isodensity
801: contours \citep[see also][]{js02,bs05-alignment}.
802: They found instead that the isopotential axis ratios are well fit
803: by the following function:
804: \begin{equation}
805: \log( \frac{b}{a} \mbox{ or } \frac{c}{a} ) =
806: \alpha \left[ \tanh\left( \gamma \log\frac{r}{r_{\alpha}} \right)
807: - 1 \right].%
808: \label{HNS equation}
809: \end{equation}
810: We have recomputed the self-consistent response of disks of varying
811: mass in a potential of this form, with the values of the parameters
812: taken from halo~G4 of \citetalias{hns07},
813: which has a very similar mass and concentration
814: to the halo used in \S~\ref{disk mass results section}.
815: The results are shown in Figure~\ref{hns figures}.
816: As noted by \citetalias{hns07},
817: the perturbation in this case contains a peak at intermediate radius
818: and is much more similar to the form of $f_{\mathrm{iso}}$ required
819: by \citetalias{hn06} to fit LSB rotation curves, although
820: unlike $f_{\mathrm{iso}}$ the
821: perturbation remains more prominent to large radius. However, in many
822: cases LSB rotation curves are only measured out to radii of a few kpc,
823: so it is less clear what the required form of $f_{\mathrm{iso}}$ is at
824: larger radii.
825: The disks have less effect on the perturbation
826: than in \S~\ref{disk mass results section}; in particular,
827: the radius inside which they circularize the potential is reduced,
828: resulting in significantly more elliptical orbits at $1$--$3$~kpc
829: than for the equivalent disks in a halo with constant
830: axis ratios.
831: Disk masses higher than $\sim 3\times 10^9~\Msun$
832: reduce the prominence of the peak in the perturbation and shift
833: it to larger radii.
834: The peak in $\fpert^{\mathrm{halo}}$ could be moved to smaller
835: radius, and therefore brought further into agreement with
836: $f_{\mathrm{iso}}$, by reducing the $r_{\alpha}$ parameter in
837: equation~(\ref{HNS equation}); however, this is unlikely to
838: be a common situation, as G4 already has by far
839: the lowest $r_{\alpha}$ value of any of the halos studied
840: by \citetalias{hns07}.
841:
842:
843: The effect of the baryonic disk on the shape of the halo
844: is not yet well understood. Simulations
845: suggest that halos containing baryonic disks
846: are less elliptical than halos composed purely
847: of dark matter, and that the circularization of the halo occurs most
848: strongly at the center \citep{kazantzidis-etal04}. If the
849: intrinsic shape of the pure dark matter halo is well described by
850: the \citetalias{hns07} form,
851: which is most elliptical at the center, baryonic processes
852: may result in a situation more similar to the constant axis
853: ratio case of \S~\ref{disk mass results section}.
854: We therefore expect that the regions in which these models
855: differ most strongly, $1$--$3$~kpc, are also the regions where the unknown
856: effect of disk formation introduces the most uncertainty into our models.
857:
858:
859: \begin{figure}
860: \epsscale{0.6}
861: \plotone{f6.eps}
862: \caption{\label{concentration figures}%
863: \textit{(a)} Magnitude of the net elliptical perturbation in the potential
864: (dashed lines) and the perturbation due to only the halo (solid lines)
865: for disks of mass $3 \times 10^9~\Msun$ in halos with
866: concentrations, $c_{200}=8$ (red), $12$ (blue), and $17$ (green).
867: The blue lines in these plots are identical to the blue lines
868: in Figure~\ref{disk mass figures}.
869: \textit{(b)} Ratio by which the input perturbation becomes
870: diluted due to the response of the disk.
871: \textit{(c)} Ellipticity of the disk isodensity contours
872: (solid lines) and of orbits within the disk (dashed lines).}
873: \end{figure}
874:
875:
876: \begin{figure}
877: \epsscale{0.6}
878: \plotone{f7.eps}
879: \caption{\label{ba figures}%
880: As in Figure~\ref{concentration figures} for halos with
881: axis ratios $b/a=0.7$ (red), $0.8$ (blue), and $0.9$ (green).}
882: \end{figure}
883:
884:
885: \begin{figure}
886: \epsscale{0.6}
887: \plotone{f8.eps}
888: \caption{\label{sys figures}%
889: As in Figure~\ref{concentration figures} for halos with
890: virial masses $M_{200}=3\times 10^{11}~\Msun$ (red),
891: $10^{12}~\Msun$ (blue), and $3\times 10^{12}~\Msun$ (green).
892: In order to facilitate
893: the comparison, the halo scale radius $r_s$ is kept constant
894: by varying $c_{200}$ in proportion to $r_{200}$, and the
895: ratio between the disk and halo mass is kept constant.}
896: \end{figure}
897:
898:
899: \begin{figure}
900: \epsscale{0.6}
901: \plotone{f9.eps}
902: \caption{\label{rd figures}%
903: As in Figure~\ref{concentration figures} for disks with
904: scale lengths $R_d=2.0$ (red), $3.0$ (blue), and $4.0$ (green).}
905: \end{figure}
906:
907: In order to investigate how other properties of the halo
908: and disk affect our results,
909: we have recalculated the results of \S~\ref{disk mass results section}
910: (where the halo axis ratio was assumed to be constant with radius)
911: for the $3 \times 10^9~\Msun$ disk while
912: varying the halo concentration,
913: the $b/a$ axis ratio,
914: the halo mass, and
915: the disk scale length.
916: In more concentrated halos (Figure~\ref{concentration figures}),
917: the strength of the perturbation due to
918: the halo, $\fpert^{\mathrm{halo}}$, is larger.
919: The disk is also less able to dilute the perturbation in more
920: concentrated halos.
921: The axis ratio of the halo (Figure~\ref{ba figures}) has a strong effect
922: on the magnitude of the perturbation but has virtually
923: no effect on the degree to which the disk dilutes the perturbation.
924: In Figure~\ref{sys figures} we compare halos of different
925: virial mass, $M_{200}$.
926: In order to facilitate comparison between systems of different
927: mass, we have kept $r_s$ constant by varying $c_{200}$ in
928: proportion to $r_{200}$ and kept the relative mass of the
929: disk and halo constant.
930: We find that the mass of the halo has little effect on the
931: relative strength of the perturbation (either before or
932: after the disk is taken into account), but that the
933: resulting disk ellipticities are higher in lower mass systems.
934: Finally, although the equilibrium shape of the potential
935: is similar regardless of the disk scale length
936: (Figure~\ref{rd figures}),
937: a greater ellipticity is required to achieve this reduction in the
938: perturbation for less concentrated disks,
939: i.e.~those with larger scale lengths.
940:
941:
942: \section{Discussion}\label{discussion section}
943: \subsection{Impact of Triaxial Halos on the Cusp/Core Problem}
944: \begin{figure}
945: \plotone{f10_color.eps}
946: %\plotone{f10-bw.eps}
947: \caption{\label{vmajmin figure}%
948: Azimuthal velocity along the halo major axis (top line) and minor
949: axis (middle line),
950: and the maximum radial velocity at each radius (bottom line)
951: for disks of varying mass in the halo of
952: \S~\ref{disk mass results section}.
953: NFW and isothermal rotation curves are shown
954: for reference in the top right panel.}
955: \end{figure}
956:
957:
958: \begin{figure}
959: \plotone{f11_color.eps}
960: %\plotone{f11-bw.eps}
961: \caption{\label{vmajmin hns figure}%
962: As in Figure~\ref{vmajmin figure}, but for the shape
963: of the halo potential found by \citetalias{hns07} for
964: their halo~G4.}
965: \end{figure}
966:
967: In Figure~\ref{vmajmin figure}, we plot the azimuthal velocities along
968: the major and minor axes of the halo within the disk plane
969: for disks of different mass in the fiducial sample halo studied
970: in \S~\ref{results section}.
971: Full analysis of the velocity fields
972: of these disks will be presented in a future paper.
973: However, we note that none of these rotation curves shows the
974: linear rise characteristic of a constant density core,
975: as expected given the dramatic difference between the form
976: of \fpert\ in these disks and the form of $f_{\mathrm{iso}}$
977: required by \citetalias{hn06}.
978:
979: If we use the radial variation of the shape of the halo potential proposed
980: by \citetalias{hns07} (Figure~\ref{vmajmin hns figure}),
981: for which $\fpert^{\mathrm{halo}}$ is much more similar to
982: $f_{\mathrm{iso}}$, we find that
983: for very low disk masses the azimuthal velocity along the
984: minor axis of the halo is characterized by a
985: much more gradual rise. Observationally, such a rotation curve might
986: be interpreted as indicating a constant-density core in the dark
987: matter halo.
988: However, for disk masses above $3 \times 10^9~\Msun$ ($0.3\%$ of the
989: virial mass of the halo and just $1.8\%$ of the baryonic
990: mass of the system) the response of the disk removes this
991: feature from the center of the rotation curve.
992: Based on these results, we conclude that simple analyses of the
993: shapes of halos are therefore not
994: sufficient to determine whether halo triaxiality can
995: reconcile LSB rotation curves with cuspy halo density profiles,
996: as suggested by \citetalias{hn06}; full analyses that take
997: into account the disk response are required. Preliminary tests on
998: simulated velocity fields constructed using the results of this paper
999: do suggest that triaxiality can produce apparent constant-density
1000: cores, in agreement with \citetalias{hn06}, but a more detailed
1001: analysis including many halos and lines of sight is needed before
1002: comparing to the observational distribution of density profile
1003: slopes \citep{simon-etal05}.
1004:
1005: We also plot the maximum radial velocity (amplitude of non-circular
1006: motions) at each radius in
1007: Figures~\ref{vmajmin figure} and \ref{vmajmin hns figure}.
1008: Although the radial velocities are much smaller than
1009: the azimuthal velocities in all but the lowest mass disks,
1010: they are at a level that can be detected in observations
1011: of two dimensional velocity fields. The magnitude of the
1012: radial motions, which reach $5$--$35~\mathrm{km~s^{-1}}$ depending
1013: on the disk mass and halo properties, are consistent with
1014: the magnitude of radial motions found by
1015: \citet{simon-etal05}, and reproduce the observed trend
1016: for the radial motions to be negligible at small
1017: radii and to only become important at larger radii.
1018:
1019:
1020: \subsection{Comparison to Observed Disk Ellipticities}
1021:
1022: It is interesting to compare the ellipticity of our model disks to
1023: observed values. Using two-dimensional velocity fields,
1024: \citet{simon-etal05} found lower limits on the orbital ellipticities
1025: ranging from $0$ up to $0.175$, similar to the orbital ellipticities
1026: in our model disks. We note that the orbital ellipticity throughout
1027: most of the disk is determined by the mass of the disk
1028: (Figure~\ref{disk mass figures}c), the ellipticity of the halo
1029: (Figure~\ref{ba figures}c), and the variation of halo ellipticity with
1030: radius (compare Figures~\ref{disk mass figures}c and \ref{hns
1031: figures}c), with very little dependence on the concentration of
1032: either the halo (Figure~\ref{concentration figures}c) or the disk
1033: (Figure~\ref{rd figures}c), or on the global mass of the system
1034: (Figure~\ref{sys figures}c). We therefore predict that galaxies with
1035: large observed ellipticities such as \object{NGC 4605} either lie in halos that
1036: are more triaxial than average near their center, or contain an
1037: unusually low fraction of their mass in their disk.
1038: The trends in intrinsic disk ellipticity that we predict may
1039: be tested with further analysis of larger kinematic samples,
1040: such as those presented by \citet{ganda-etal06}.
1041:
1042: \citet{ryden06} found that the distribution of isophotal shapes
1043: of galaxies in the 2MASS Large Galaxy Atlas \citep{2mass-lga},
1044: as measured in the near-infrared $K_s$ band (which is a good tracer
1045: of the stellar disk mass), is well fit if
1046: the intrinsic disk ellipticity distribution is a truncated Gaussian
1047: centered at $0.01$ with a width of $\sigma=0.26$. This corresponds
1048: to a median ellipticity of $0.18$, with 68\%\ of disks having
1049: ellipticities $0.05 \le \epsdisk \le 0.37$. For the disk
1050: parameters we have studied, the
1051: $18.8~\mathrm{mag~arcsec^{-2}}$ isophote at which her shapes
1052: were measured corresponds to radii
1053: of between $1.5$ and $8.5~\mathrm{kpc}$. Our models naturally
1054: produce disk ellipticities in this range at these radii,
1055: although the highest
1056: ellipticities can only be produced by our least massive disks in
1057: our most elliptical potentials.
1058:
1059: Finally, we note that our models only take into account
1060: ellipticity in the disk induced by the dark matter halo.
1061: Central regions of the disk, which have high surface density
1062: and sit in an axisymmetric potential, may be unstable to
1063: bar formation \citep{bs06}. This can induce additional ellipticity
1064: to the kinematic and photometric properties of disk galaxies
1065: \citep[e.g.,][]{valenzuela-etal07}.
1066:
1067:
1068: \section{Conclusions}\label{conclusions section}
1069: We have presented a
1070: computationally efficient method to self-consistently determine the dynamics
1071: of massive disks in triaxial dark matter halos. Our work extends
1072: the study of \citetalias{jog00} by allowing the perturbation
1073: to the potential to vary with radius in an appropriate manner and by
1074: allowing the ellipticity of the disk to vary with radius
1075: self-consistently. These improvements result in qualitatively
1076: different behavior for the ellipticity of disks at small radii:
1077: \citetalias{jog00} found that disks counteract the halo
1078: ellipticity most strongly at $1.42~R_d$
1079: and have a negligible effect at small radii; in contrast, we find
1080: that the effect of the disk increases monotonically
1081: to small radii, completely circularizing the potential
1082: in the innermost regions.
1083:
1084: This self-consistent radially-varying response of the disk to the halo
1085: perturbation must be taken into account when comparing the observed
1086: kinematic and photometric properties of galactic disks to those
1087: expected in triaxial dark matter halos, particularly for comparisons
1088: at small radii. When this response is calculated for plausible halo
1089: values, model disks have ellipticities consistent with those
1090: determined from observations of velocity fields and from isophotal
1091: axis ratio distributions. We also find that the radial variation of
1092: the halo axis ratios has a significant impact on the disk structure.
1093: Halos with axis ratios that vary with radius as suggested by
1094: cosmological simulations produce much more elliptical orbits
1095: in the inner disk than do halos with constant axis ratios,
1096: resulting in potential perturbations similar to
1097: the perturbation required to create apparent cores in galaxy density
1098: profiles.
1099: Further analysis exploring in detail the conditions under
1100: which core-like rotation curves might be obtained will be necessary to
1101: determine if halo triaxiality can resolve the cusp/core problem.
1102:
1103:
1104: \acknowledgements
1105: JB thanks the Australian Research Council for financial support. JDS
1106: acknowledges the support of a Millikan Fellowship provided by Caltech.
1107: BKG and CP gratefully acknowledge the support of the Australian
1108: Research Council supported ``Commonwealth Cosmology Initiative'',
1109: DP 0665574.
1110: We thank Milo{\v s} Milosavljevi{\'c} and Darren Croton
1111: for helpful conversations, and the referee, Chanda Jog,
1112: for a very useful report.
1113:
1114: \bibliography{../../masterref.bib}
1115:
1116: \appendix
1117: \section{Second-order terms and m=4 distortions}\label{second order terms section}
1118: We have assumed that the halo perturbation and the
1119: induced distortion in the disk are
1120: completely described by the $m=2$ mode.
1121: This a direct consequence of only including terms linear in
1122: the small quantities \fpert, \epsdisk, and their derivatives.
1123:
1124: The validity of this assumption can be tested by evaluating
1125: the magnitude of the second-order terms that contribute to the
1126: $m=4$ distortion in the disk. If we expand the potential as
1127: \begin{equation}
1128: \Phi(R,\theta) = \Phi_0(R) \left(1 + f_2(R) \cos 2\theta +
1129: f_4(R) \cos 4\theta\right),
1130: \end{equation}
1131: the disk surface density as
1132: \begin {equation}
1133: \Sigma(R,\theta) = \Sigma_0 \exp \left[ - \frac{R}{R_d}
1134: \left( 1 - \frac{\epsilon_2(R)}{2} \cos 2\theta
1135: - \frac{\epsilon_4(R)}{2} \cos 4\theta \right) \right],
1136: \end{equation}
1137: and include all second-order terms, then the $m=4$ distortion in the
1138: disk, $\epsilon_4$, depends on terms of order
1139: $f_4$, $f_2\, \epsilon_2$, and
1140: $R\, f_2\, \diffd\epsilon_2/\diffd R$.%
1141: \footnote{As in the case of the $m=2$ mode, there is also
1142: a negligible term of order $R\, \diffd(f_4\, a_{14})/\diffd R$.}
1143: \begin{figure}
1144: \plotone{f12.eps}
1145: \caption{\label{eps4 figure}%
1146: Magnitude of terms contributing to the $m=4$ distortion
1147: in the disk compared to $\epsilon_2$,
1148: the magnitude of the calculated $m=2$ distortion
1149: for a disk of mass $3 \times 10^9~\Msun$ in the fiducial halo
1150: of \S~\ref{disk mass results section}.}
1151: \end{figure}
1152: Figure~\ref{eps4 figure} compares
1153: the magnitude of these terms to the $m=2$ ellipticity for the
1154: $3 \times 10^9~\Msun$ disk in the fiducial halo of
1155: \S~\ref{disk mass results section}.
1156: The higher-order terms are more than two orders of magnitude
1157: smaller than the first-order terms over most of the disk,
1158: and are also negligible in the central region where the first order terms
1159: vanish,
1160: validating our use of linear perturbation theory.
1161:
1162:
1163:
1164: \end{document}
1165:
1166: