1: \documentclass{edp-conf}
2:
3: \usepackage{graphics}
4: \usepackage[OT1]{fontenc}
5:
6:
7: \newcommand{\vi}{\vec{v}}
8: \newcommand{\vxi}{\vec{\xi}}
9: \newcommand{\lp}{ \left(}
10: \newcommand{\rp}{ \right)}
11: \newcommand{\na}{ \vec{\nabla} }
12: \newcommand{\disp}[1]{\displaystyle #1}
13: \newcommand{\er}{\vec{e}_r}
14: \newcommand{\noi}{ \noindent }
15: \newcommand{\dnr}[1]{\frac{d #1}{dr}}
16: \newcommand{\ddnr}[1]{\frac{d^2 #1}{dr^2}}
17: \newcommand{\dnz}[1]{\frac{d #1}{d\zeta}}
18: \newcommand{\ddnz}[1]{\frac{d^2 #1}{d\zeta^2}}
19: \newcommand{\dnxi}[1]{\frac{d #1}{d\xi}}
20: \newcommand{\ddnxi}[1]{\frac{d^2 #1}{d\xi^2}}
21: \def\Ln{\mathop{\hbox{ln}}\nolimits}
22: \newcommand{\beq}{\begin{equation}}
23: \newcommand{\eeq}{\end{equation}}
24: \newcommand{\dr}[1]{\frac{\partial #1}{\partial r}}
25: \newcommand{\ltex}[1]{\quad \hbox{#1} \quad}
26: \newcommand{\llp}{ \ell(\ell+1)}
27: \newcommand{\lc}{ \left[}
28: \newcommand{\rc}{ \right]}
29: \newcommand{\greq}{\begin{equation}\left\{ \begin{array}{l}}
30: \newcommand{\egreq}{\end{array}\right. \end{equation}}
31: \newcommand{\uu}{ u^\ell_m }
32: \newcommand{\vv}{ v^\ell_m }
33: \newcommand{\YL}{ Y^m_\ell }
34: \newcommand{\YTL}{\hat{Y}^m_\ell}
35: \newcommand{\YTLP}{\hat{Y}^m_{\ell '}}
36: \newcommand{\DYTLP}{\frac{\partial \hat{Y}^m_{\ell '}}{\partial \theta}}
37: \newcommand{\lpp}{ \ell \ell '}
38: \newcommand{\lkk}{ k k '}
39: \newcommand{\dnx}[1]{\frac{d #1}{dx}}
40: \newcommand{\eeqn}[1]{\label{#1}\end{equation}}
41: \newcommand{\od}[1]{\mbox{${\cal O}(#1)$}}
42: \newcommand{\eq}[1]{(\ref{#1})}
43: \newcommand{\beqa}{\begin{eqnarray*}}
44: \newcommand{\eeqa}{\end{eqnarray*}}
45:
46: \begin{document}
47:
48: \TitreGlobal{SF2A 2001}
49:
50: \title{Acoustic modes in spheroidal cavities}
51:
52: \author{F. Ligni\`eres}
53: \address{
54: Laboratoire d'Astrophysique, CNRS UMR 5572, Obs. Midi-Pyr\'en\'ees,
55: 57 avenue d'Azereix, 65008 Tarbes Cedex, France}
56: \author{M. Rieutord}
57: \address{
58: Laboratoire d'Astrophysique de Toulouse,
59: Observatoire Midi-Pyr\'en\'ees, 14 avenue E. Belin, 31400 Toulouse,
60: France and Institut Universitaire de France}
61: \author{L. Valdettaro}
62: \address{
63: Dipartimento di Matematica, Politecnico di Milano, Piazza L.
64: da Vinci, 32, 20133 Milano, Italy
65: }
66:
67: \runningtitle{Acoustic modes in spheroidal cavities}
68:
69: \maketitle
70:
71: \begin{abstract}
72: Oscillation modes of rapidly rotating stars have not yet been calculated
73: with precision, rotational effects being generally approximated by
74: perturbation methods. We are developing a numerical method able to
75: account for the deformation of the star by the centrifugal force and,
76: as a first step, we determined the acoustic modes of a uniform density
77: spheroid and studied how the spheroid flatness affects these modes.
78: \end{abstract}
79:
80: \section{Introduction}
81:
82: The ratio of centrifugal and gravity forces being generally small in
83: stars, the effect of rotation on gravito-acoustic stellar oscillations
84: has been mostly studied with perturbation methods. Although fully
85: justified in the context of helioseismology, this approach might not be
86: accurate enough for rapidly rotating stars when the difference between the
87: correct and approximated frequency becomes comparable with the accuracy
88: of frequency measurements. We thus expect that new non-perturbative
89: approaches will be necessary in the context of future asteroseismology
90: missions (Mons, Corot, Eddington).
91:
92: Such non-perturbative calculations should include the effects of the
93: Coriolis and centrifugal forces on the wave motions as well as the
94: effect of the centrifugal force on the equilibrium state of the star.
95: This latter effect breaks down the spherical symmetry of the cavity
96: within which acoustic modes resonate and this introduces mathematical
97: difficulties since the related eigenvalue problem is no longer fully
98: separable (except in the particular case of spheroids).
99:
100: In this paper, we present a numerical method able to solve this type of
101: problem. Before it is applied to a realistic stellar model, we tested
102: it in a simpler case. We choose the homogeneous spheroid because an
103: alternative method exists in this case. Let also recall that stationary
104: solutions of a self-graviting and uniformly rotating gas of constant
105: density are spheroids (the so-called Maclaurin spheroids). Thus acoustic
106: modes of homogeneous spheroids can be viewed as the oscillation modes
107: of Maclaurin spheroids in the high frequency limit (in order to neglect
108: the effect of Coriolis force on the oscillatory motions) and within the
109: Cowling approximation (to neglect the effect of gravity on the modes).
110: In the next sections, the formalism is presented and the result of the
111: test as well as a first analysis of the flatness effects on the frequency
112: spectrum are given.
113:
114: \section{The formalism}
115:
116: Oscillatory motions of small amplitude in a perfect fluid of uniform
117: density are governed by the Helmholtz equation,
118:
119: \begin{equation}
120: \Delta \hat{\psi} + k^2 \hat{\psi} = 0,
121: \label{eqdl2}
122: \end{equation}
123: \noindent where $\psi = \hat{\psi} \exp(i \omega t)$ is the velocity
124: potential (${\bf u} = \nabla \psi$) and $k = \omega/c_s$, the ratio of
125: the mode frequency, $\omega$, and the sound speed, $c_s$. The boundary
126: condition at the surface of the cavity is, ${\bf n}\cdot\nabla \hat{\psi}
127: = 0,$ where ${\bf n}$ is a vector normal to the surface.
128:
129: Expect for the special cases of spherical or spheroidal cavities this
130: eigenvalue problem is not fully separable. Then, for arbitrary axially
131: symmetric surfaces, a 2D eigenvalue problem must be solved numerically.
132: We first note that, for numerical reasons, it is preferable to apply
133: the boundary conditions on a coordinate surface. A coordinate
134: system must thus be chosen so that the surface be a surface of
135: coordinate ($\zeta = cste$). That is, if $r= S(\theta)$ describes
136: the surface, suitable coordinates are $(\zeta = f(\zeta, \theta),
137: \theta, \phi)$ where for some value $\zeta_0$ the function $f$ verifies
138: $r=f(\zeta_0,\theta)=S(\theta)$. In this expressions $(r, \theta, \phi)$
139: are the usual spherical coordinates. Following Bonazzola et al. (1998),
140: $f$ is then specified so that the regularity conditions at the center
141: have a simple form. For spheroids, we used,
142:
143: \begin{equation}
144: r = f(\zeta, \theta) = R_e \left[\zeta + (3 \zeta^4 - 2 \zeta^6)\left(\frac{1}{\sqrt{1 + g(\epsilon) \cos^2\!\theta}} - 1\right)
145: \right]
146: \end{equation}
147: \noindent
148: where $R_e$ and $R_p$ are respectively the semi-major axis and the
149: semi-minor axis, $\epsilon = 1 - R_p / R_e$ is the flatness and
150: $g(\epsilon) = \epsilon(2-\epsilon)/(1-\epsilon)^2$. The eigenvalue
151: problem is written with these new coordinates. By expanding the
152: solution $\hat{\psi}(\zeta,\theta,\phi)$ on spherical harmonics and
153: projecting the equations on each spherical harmonic $\YL(\theta,\phi)$,
154: one obtains for each value of the azimuthal number $m$ two sets of ODE
155: of the variable $\zeta$ coupling the coefficients of spherical harmonic
156: expansion corresponding respectively to odd and even degree numbers,
157: that is ($\hat{\psi}^{m+2k}_m (\zeta), \; 0 \leq k < +\infty$) and
158: $(\hat{\psi}^{m+2k+1}_m(\zeta), \; 0 \leq k < +\infty)$. Then following
159: a method developed by Rieutord and Valdettaro (1997) after Orszag (1971),
160: the ODE are discretized on the Gauss-Lobato grid associated with Chebyshev
161: polynomials. A set of linear equations dependent of the parameter $k$
162: results and non-trivial solutions of this system are found using either
163: the QZ algorithm or the Arnoldi-Chebyshev algorithm.
164:
165: To test this general method, we considered a spheroidal cavity as a simple
166: alternative method exists in this case. This is because the eigenvalue
167: problem is separable using the so-called oblate spheroidal coordinates
168: $(\xi, \eta, \phi)$ defined as $(x = a\cosh\xi\sin\eta\sin\phi, \; y =
169: a\cosh\xi\sin\eta\cos\phi, \; z = a \sinh\xi\cos\eta)$, where $0\leq \xi
170: <+ \infty$, $0\leq \eta \leq \pi$ et $0\leq \phi \leq 2 \pi$. The package
171: Linear Solver Builder developed by Rieutord and Valdettaro was also used
172: to solve the eigenvalue problem written in this coordinate system.
173:
174: Finally, we tested our results in the range of small flatness using
175: a perturbation approach. Following the method described in Gough et
176: al. (1990), we calculated analytically the first order effect of the
177: flatness on the mode frequencies. It reads,
178:
179: \begin{equation}
180: \tilde{\omega}_{n\ell m} = \tilde{\omega}^0_{n\ell} +
181: \frac{2 \epsilon \tilde{\omega}^0_{n\ell}}{3 (2\ell -1)(2\ell +3)}\left(
182: \ell(\ell +1) - 3 m^2 + 3 \frac{\ell(\ell +1) -3m^2}
183: { {\left({\tilde{\omega}}^{0}_{n\ell}\right)}^2 - \ell(\ell +1)} \right),
184: \label{eqdl}
185: \end{equation}
186: \noi where $\tilde{\omega}^0_{n\ell} = {\omega}^0_{n\ell} R/ c_s$
187: is the dimensionless eigenfrequency corresponding to the sphere of the
188: same volume, $R = R_e \left( 1 - \epsilon \right)^{1/3}$. It is given as
189: the $n^{\mbox{th}}$ root (by growing order) of the following equation,
190: $(\ell+1) J_{\ell+1/2}(x) - x J_{\ell-1/2}(x) = 0, \; 1 \leq \ell <
191: +\infty$, where $J_{\ell}$ are Bessel functions.
192:
193: \section{Results}
194:
195: We found that the general method and the method specific
196: to spheroidal cavities provide the same frequencies with
197: a high level of accuracy for arbitrary value of the flatness
198: between $0$ and $0.5$.
199: Moreover, as the flatness goes to zero, the frequencies
200: converge towards the values given by the asymptotic analysis (Eq. \ref{eqdl}).
201: We are thus encouraged in using the general method described in this paper
202: for future pulsation
203: models which shall take
204: into account the deformation of the star equilibrium
205: by the centrifugal force.
206:
207: We also studied
208: how the flatness of the spherical cavity affect the
209: frequency of acoustic modes.
210: The first point is to understand why, for increased flatness,
211: the frequency of some
212: modes increases while the frequency of others decreases.
213: Simple arguments concerning the region of mode propagation
214: can explain these behaviors.
215: For large degree ($\ell \gg 1$) quantitative results can even
216: be deduced from such arguments: Large degree axi-symmetric modes ($\ell \gg 1,
217: m=0$) propagate in meridional planes just below the cavity surface.
218: Therefore, the characteristic length
219: involved in their quantization
220: condition shall be the perimeter of the circle, $2 \pi R$, for the sphere,
221: and the perimeter of the ellipse
222: for the spheroid.
223: By contrast sectorial modes ($m=\ell$) are confined towards
224: the equator so that the characteristic length is the perimeter
225: of the equatorial circle that is $2 \pi R$ for the sphere and $2\pi R_e$
226: for the spheroid.
227: The increase or decrease of the modes frequency should
228: then be determined by the ratio of these
229: characteristic lengths.
230: This interpretation is consistent with the first order asymptotic formula
231: (Eq. \ref{eqdl}) since in the large $\ell$ limit, the frequency ratio
232: $\tilde{\omega}_{n\ell m} / \tilde{\omega}^0_{n\ell}$
233: equals the ratio of the characteristic lengths.
234:
235: As a consequence of the differential effect of flatness
236: on frequencies, modes with originally distinct frequency
237: tends to reach identical frequencies when flatness increases.
238: However, this is not
239: allowed for the coupled modes and avoided crossing occur.
240: As all the modes having identical azimuthal number
241: and degree numbers of the same parity are coupled (see above),
242: avoided crossings occur at a strong rate in such a way
243: that
244: for non negligible
245: flatness most of the modes have
246: experienced many avoided crossing.
247:
248: Another interest of the present calculation
249: is to assess the limit of validity of the perturbative approach.
250: Note first that, for small rotation rates, the flatness of MacLaurin spheroids
251: is proportional to the ratio of centrifugal and gravity forces.
252: Thus first order effects in terms of flatness correspond to second order effect
253: in terms of the rotation rate.
254: We found that the difference between the correct and the approximated frequency
255: depends on the type of modes. But it can be relatively important.
256: For example, at $\epsilon = 0.1$, the relative difference
257: for axisymmetric modes of degree $\ell = 1$ grows with
258: the radial order and for $1 \leq n \leq 6$ it is comprised between
259: $0.01$ and $0.03$.
260: If we extrapolate such relative differences to a rapidly rotating $\delta$ Scuti
261: oscillating around 200 $\mu$ Hz, we find a frequency difference of 1 to 3 $\mu$ Hz that
262: is one order of magnitude larger than the accuracy planned for
263: Corot mission.
264: Although a firm conclusion should await stellar oscillation
265: models, the present estimate points towards the necessity of
266: using non-perturbative method for spatial seismology of
267: rapidly rotating stars.
268:
269: We also noticed the apparition of new types of modes
270: in the spheroidal geometry.
271: They are easily identified
272: in the context of ray theory of axi-symmetric modes.
273: Whereas for spherical cavity, all caustics are concentric spheres,
274: two type of caustics, concentric spheroids and hyperboloids,
275: are possible within a spheroidal cavity.
276: In this last case, the ray does not propagate in the equatorial regions
277: comprised between the surface boundary and the hyperbolic caustic.
278: With our calculation we effectively found modes which show
279: hyperbolic like regions of very low amplitude around the equator.
280:
281: \section{Future work}
282:
283: The next step is to calculate the oscillations
284: of a rotating polytropic star including the effect of Coriolis
285: and centrifugal forces on the motions. Among the new features
286: arising when considering these more realistic stellar models
287: is the chaos of rays
288: dynamics within non-spherical and non-spheroidal cavities.
289: Astrophysical consequences of this
290: chaotic behavior shall be investigated.
291:
292: \begin{thebibliography}{4}
293: \bibitem{BGM} Bonazzola, S., Gourgoulhon, E., Marck, J-A. 1998, Phys. Rev. D, 58, 104020
294: \bibitem{GT} Gough, D.O., Thompson, M.J., 1990, Mon. Not. R. astr. Soc., 242, 25
295: \bibitem{O} Orszag, S.A., 1971, J. Fluid Mech., 50, 689
296: \bibitem{RV} Rieutord, M., Valdettaro, L. 1997, J. Fluid Mech., 341, 77
297: \end{thebibliography}
298: \end{document}
299: