1: %\documentclass[preprintnumbers,amsmath,amssymb]{revtex4}

2: \documentclass[12pt]{iopart}

3: %\documentclass[12pt]{article}

4: 

5: \usepackage[dvips]{graphicx}

6: %\usepackage{amsmath,amsfonts,amssymb}

7: 

8: \begin{document}

9: 

10: \title{General Relativistic Galaxy Rotation Curves: Implications for Dark Matter Distribution}

11: 

12: \author{Dylan Menzies and Grant J. Mathews}

13: 

14: \address{Department of Physics, Center for Astrophysics, University of Notre Dame, Notre Dame, IN 46556  email:  dmenzies@nd.edu,  gmathews@nd.edu}

15: 

16: 

17: \begin{abstract}

18: It has recently been suggested  that observed galaxy rotation curves can be accounted for by general relativity without recourse to dark-matter halos.  Good fits have been produced to observed galatic rotation curves using this model.  We show that the implied total mass is infinite, adding to the evidence opposing the hypothesis.

19: \end{abstract}

20: 

21: 

22: \maketitle

23: 

24: \section{Introduction}

25: Understanding the nature and origin of the apparent dark matter responsible for the bulk  of the gravitational mass of galaxies and clusters is one of the most important challenges in modern cosmology.

26: Although many candidates have been proposed for the required weakly interacting cold dark matter \cite{Baltz04},

27: to date there has been no convincing detection of a suitable particle candidate.  For this reason there is considerable interest in a recent proposal \cite{CT} that accurate modeling of galactic dynamics using general relativity is consistent with observed flat rotation curves and luminous mass without the need for additional dark matter. In this scenario, the authors assume a cylindrical stationary space-time containing pressureless dust and solve the weak-field Einstein equation to first order in the gravitational constant $G$ to obtain the associated velocity distribution as described below.

28: 

29: 

30: At first glance, this would seem to be a compelling solution to the dark matter problem.

31: However, objections have already been raised by several authors, on a various grounds.  These include   the conditions at $z=0$ \cite{korz,VL}, general approximation principles \cite{garf}, and specific analysis of the approximations used \cite{cross}. The original authors have responded to some of these criticisms \cite{response}.   In this paper, we suggest yet another possible problem with this interpretation of the dark-matter problem.  We reanalyze the galactic rotation curve as in \cite{CT} and find that

32: the model, as applied, contains an implied large (infinite) mass distribution extending beyond the

33: observed galactic disk, and hence at best has simply moved the dark-matter to the outer Galaxy, and

34: more likely, the model is ruled out on energetic grounds.

35: 

36: \section{The Cooperstock and Tieu Model}

37: Our goal is to analyze the implied mass distribution in the Cooperstock \& Tieu \cite{CT}  model.

38: In order to deduce this we first summarize their approach.

39: In their analysis of  rotation curves, the galaxy to be described is treated as

40: a uniformly rotating axisymmetric system containing  pressureless fluid.

41: The general stationary axially symmetric

42: metric is chosen of  the form

43: \begin{equation}

44: 	ds^2 =	-e^{\nu-w}( udz^2+dr^2)

45: 		-r^2 e^{-w} d\phi^2+e^w(cdt-Nd\phi)^2~~,

46: \label{Eq1}

47: \end{equation}

48: where $u$, $\nu$, $w$ and $N$ are functions of cylindrical polar

49: coordinates $r$, $z$, and the metric coefficient

50: $u$ is taken to be unity.  In a co-moving frame the

51: the fluid four-velocity is just

52: \begin{equation}

53: 	U^i = {\delta}_0^i ~~.

54: \label{Eq2}

55: \end{equation}

56:  Making  a local

57: ($r$, $z$ held fixed) transformation to

58: \begin{equation}

59: 	\bar{\phi} = \phi + \omega(r,z) \,t~~,

60: \label{Eq3}

61: \end{equation}

62:  diagonalizes the metric and,  in the weak field limit, leads to simple approximations for

63:   the local angular velocity $\omega$ and the tangential

64: velocity $V$:

65: \begin{equation}

66: 	\omega

67: 	= \frac{Nce^w}{r^2e^{-w}-N^2e^w}

68: 	   \approx  \frac{Nc}{r^2}~~,  \\

69: 	V =\omega r~~.

70: \label{Eq4}

71: \end{equation}

72: 

73: The Einstein field equations to first order in the gravitational constant $G$

74:  are easily combined to yield a Poisson-like equation for $\omega$.

75: \begin{equation}

76: 	\nabla^2 w +\frac{N_r^2+N_z^2}{r^2}=\frac{8\pi G\rho}{c^2}~~,

77: \end{equation}

78: where

79: \begin{equation}

80: 	\nabla^2 w \equiv w_{rr} + w_{zz} + \frac{w_r}{r}~~.

81: \end{equation}

82: 

83: The geodesic

84: equation

85: \begin{equation}

86: 	\frac{dU^i}{ds} +\Gamma^i_{kl}U^k U^l=0

87: \end{equation}

88: together with the adopted metric and four-velocity

89:  (\ref{Eq1}) and (\ref{Eq2}) simply becomes

90: \begin{equation}

91: 	w_r=w_z=0~~,

92: \end{equation}

93: and hence

94: \begin{equation}

95: 	\nabla^2 w=0~~,

96: \end{equation}

97: within the fluid.  The interior equations for $\rho$ and $N$ then become:

98: \begin{equation}

99: 	N_{rr} + N_{zz} - \frac{N_r}{r} =0

100: \label{Eq9a}

101: \end{equation}

102: \begin{equation}

103: 		\frac{N_r^2 + N_z^2}{r^2} = \frac{8{\pi}G\rho}{c^2}~~.

104: \label{Eq9b}

105: \end{equation}

106: Eq. (\ref{Eq9a})

107: can be expressed as

108: \begin{equation}

109: 	\nabla^2\Phi =0

110: \label{Eq10}

111: \end{equation}

112: where

113: \begin{equation}

114: 	\Phi \equiv \int\frac{N}{r}dr~~.

115: \label{Eq10a}

116: \end{equation}

117: Hence,  flat-space harmonic functions $\Phi$ become  the generators of the

118: axially symmetric stationary pressure-free weak fields.

119:  Using (\ref{Eq4}) and (\ref{Eq10a}),  the

120: tangential velocity distribution becomes

121: \begin{eqnarray}

122: 	V &=& c\frac{N}{r} \\

123: 	&=& c\frac{\partial {\Phi}}{\partial{r}}~~.

124: \end{eqnarray}

125: Since the field equation for $\rho$ is non-linear, the galactic rotation is deduced

126: by first finding the required generating

127: potential $\Phi$ and from this the appropriate function $N$

128: for the galaxy is deduced. Then Eq.~(\ref{Eq9b})

129: yields the density distribution.

130: 

131: Each galaxy analyzed requires its own composing elements to

132: build the generating potential. A separation of variables yields a solution for $\Phi$ in terms of Bessel functions $J_n(kr)$

133: \begin{equation}

134: 	\Phi = Ce^{-k\mid z \mid}J_0(kr)

135: \end{equation}

136: where $C$

137: is an arbitrary constant. The general solution is expressed as a linear superposition

138: \begin{equation}

139: 	\Phi = \sum_{n}C_ne^{-k_n |z|}J_0(k_nr)

140: \label{Eq13}

141: \end{equation}

142: with $n \sim 10$ chosen appropriately for the desired level of accuracy.

143: From (\ref{Eq13}) and (\ref{Eq4}), the tangential velocity is

144: \begin{equation}

145: 	V= -c\sum_{n} k_n C_n e^{-k_n |z|}J_1(k_nr)~~.

146: \end{equation}

147: 

148: This expression was used in \cite{CT} to fit the rotation curve for the Milky Way.

149: In this paper we analyze this fit and show that any fit implies a large mass at larger radii.

150: Hence, this approach does not solve the dark matter problem, but has merely moved to

151: dark matter to large radii. Worse, we now have an excess mass problem!

152: 

153: 

154: \section{Interpretation of the Model Fits}

155: 

156: Figure \ref{fig:vel500} reproduces the velocity curve fit for the Milky Way given in \cite{CT}, but extended to larger radius. The large oscillations in velocity although alarming, are rendered harmless if the density is sufficiently low. Figure \ref{fig:density100} shows the extended density profile. It would be tempting to assume from this  that the density could be cutoff at the limits of the observed galaxy. However, we show that for any velocity fit the total mass is infinite, in contrast  to the original claim of finite integrated mass. In fact, smoothed out,  the enclosed mass rises linearly with radius.

157: 

158: \begin{figure}

159: \begin{center}

160: \includegraphics[width=8cm]{vel500.eps}

161: \caption{\label{fig:vel500} Velocity $V \ (m/s)$ as a function of radius $r \ (kpc)$}

162: \end{center}

163: \end{figure}

164: 

165: \begin{figure}

166: \begin{center}

167: \includegraphics[width=10cm]{density100.eps}

168: \caption{\label{fig:density100} Density $\rho \ (kg m^{-3})$ as a function of radius $r \ (kpc)$}

169: \end{center}

170: \end{figure}

171: 

172: 

173: \section{Integrating the mass}

174: To obtain the total implied galactic mass, consider a model with a single value of $k$,

175: \begin{equation}

176: N = r e^{-kz} J_1(kr) ~~.

177: \end{equation}

178: Eq (\ref{Eq9b}) written for density is

179: \begin{equation}

180: \label{eq:density}

181: \rho = \frac{c^2}{8\pi G}\left(\frac{N_r^2+N_z^2}{r^2}\right) ~~.

182: \end{equation}

183: In this case

184: \begin{equation}

185: \rho = e^{-2kz} (J_0(kr)^2 + J_1(kr)^2)~~ .

186: \end{equation}

187: We note $r \rho \rightarrow({2}/{\pi} ) \exp{\{-2z\}}$ as $r \rightarrow \infty$, illustrated in Figure \ref{fig:rdensityk}.

188: \begin{figure}

189: \begin{center}

190: \includegraphics[width=10cm]{rdensityk.eps}

191: \caption{\label{fig:rdensityk} $r \rho$ as a function of radius $r \ (kpc)$, $z = 0$}

192: \end{center}

193: \end{figure}

194: The mass element is

195: \begin{equation}

196: dm = \rho dr dz \, r d\theta~~.

197: \end{equation}

198: So the mass enclosed within  cylindrical coordinate r is

199: \begin{eqnarray}

200: \label{eq:M}

201: M(r) &=& 2 \pi \int dr dz \, r\rho \\

202:  &\rightarrow& \biggl[2 \pi \biggl(\frac{2}{\pi}\biggr) \int dz e^{-kz} \biggr] r~~ .

203: \end{eqnarray}

204: In this case $M(r)$ clearly rises linearly.

205: 

206: Now we look at the general case. Taking our cue from Eq.(\ref{eq:M}), $r \rho$ is plotted against $r$ for the Milky Way model using two different scales of extended range, shown in Figures \ref{fig:rdensity100}, \ref{fig:rdensity1000}.

207: \begin{figure}

208: \begin{center}

209: \includegraphics[width=10cm]{rdensity100.eps}

210: \caption{\label{fig:rdensity100} $r \rho$ as a function of radius $r \ (kpc)$, $z = 0$}

211: \end{center}

212: \end{figure}

213: \begin{figure}

214: \begin{center}

215: \includegraphics[width=10cm]{rdensity1000.eps}

216: \caption{\label{fig:rdensity1000} $r \rho$ as a function of radius $r \ (kpc)$, $z = 0$}

217: \end{center}

218: \end{figure}

219: It appears that $r \rho$ oscillates around a positive average value. If the $z$ dependence were independent of $r$ like that in the single $k$ case, this would imply that $M(r)$ tends to rise linearly in the limit of large $r$. The following analysis shows this is the case, even though the $z$ dependence is not simple.

220: 

221: If we can find $\alpha, \beta>0$ such that for large enough $r$

222: \begin{equation}

223: \frac{\int dz r\rho(r,z)}{r\rho(r,0)} > \alpha

224: \end{equation}

225: and

226: \begin{equation}

227: \int_0^r r\rho(r,0) > \beta r

228: \end{equation}

229: then $M(r)> \alpha\beta r$ for large enough $r$. To find $\alpha$, for fixed $r$ define

230: \begin{equation}

231: f(z) = \frac{r\rho(r,z)}{r\rho(r,0)} .

232: \end{equation}

233: From (\ref{eq:density}),

234: \begin{equation}

235: f(z) = (\sum_k a_k e^{-k z})^2 + (\sum_k b_k e^{-k z})^2 ,

236: \end{equation}

237: with coefficients $a_k, b_k$ determined by the model. Note $f(0) = 1$. Define $\alpha$ as the minimum over all sets ${a_k, b_k}$, in any model, of the integral,

238: \begin{equation}

239: \alpha = \min_{\{a_k, b_k\}} \int dz f(z) .

240: \end{equation}

241: $\alpha > 0$ since otherwise $f(z) \ge 0$ and continuity of $f(z)$ imply $f(0)=0$, a contradiction.

242: \\

243: \\

244: To find $\beta$ first find the form of the density for $z=0$ using (\ref{eq:density}),

245: \begin{equation}

246: \rho(r,0) = \left[ \left(\sum_k \left( r a_k J_0(kr) + c_k J_1(kr)\right)\right)^2 + \left(\sum_k r b_k J_1(kr)\right)^2 \right]/r^2 ,

247: \end{equation}

248: 

249: with coefficients determined by the model. So as $r \rightarrow \infty$

250: \begin{eqnarray}

251: r\rho &\rightarrow& r\left[ \left( \sum_k a_k J_0(kr) \right) ^2 + \left(\sum_k b_k J_1(kr)\right)^2 \right] \\

252:       &\rightarrow& \left( \sum_k  a_k \sin{kr} \right)^2 + \left( \sum_k  b_k \cos{kr} \right)^2 ,

253: \end{eqnarray}

254: with a re-scaling of the coefficients. So the integral

255: \begin{eqnarray}

256: \int dr r\rho = \sum_k \int_0^r dr (a_k \sin{kr})^2 + (b_k \cos{kr})^2 \\

257:  +  \sum_{k_1 \ne k_2} \int_0^r dr \left( a_{k_1}a_{k_2} \sin{k_1 r} \sin{k_2 r} + b_{k_1}b_{k_2} \cos{k_1 r} \cos{k_2 r} \right) \nonumber

258: \end{eqnarray}

259: The first sum consists of a piece linear in $r$ plus an oscillating piece, while the second sum is purely oscillating. The oscillations are all bounded, so overall the required $\beta>0$ exists, and $M(r)$ behaves as predicted.

260: 

261: \section {Conclusion}

262: Despite the close fit to observed velocity curves over the galatic range, we find that the total enclosed mass in the model of Cooperstock and Tieu rises as a linearly increasing function of radius. This indicates that the model has not removed the dark matter problem, but has merely moved the mass to large radius.  Moreover, the implied infinite mass, and large velocity fields suggest that this interpretation is probably not realistic.

263: 

264: 

265: \vskip .2 in

266: Discussions with Bill McGlinn are gratefully acknowledged. Work at the University of Notre Dame supported by the U.S. Department of Energy under research grant DE-FG02-95-ER40934.

267: 

268: \section*{References}

269: \begin{thebibliography}{99}

270: 

271: \bibitem{Ostriker03} Ostriker, J. P. and Steinhardt, P. 2003, Science, 300,  1909

272: \bibitem{Baltz04} Baltz, E. A. 2004, astro-ph/0412170

273: \bibitem{CT} Cooperstock, F.I. and Tieu, S. 2005, astro-ph/0507619

274: \bibitem{korz} Korzynski, M. 2005, astro-ph/0508377

275: \bibitem{VL} Vogt, B. and Letelier, 2005 P.S., astro-ph/0510750

276: \bibitem{garf} Garfinkle, D. 2005,  Class.Quant.Grav. 23 (2006) 1391, gr-qc/0511082

277: \bibitem{cross} Cross, D.J. 2006, astro-ph/0601191

278: \bibitem{response} Cooperstock, F.I. and Tieu, S. 2005, astro-ph/0512048

279: 

280: \end{thebibliography}

281: 

282: \end{document}

283: 

284: