1: %\documentclass[preprint,epsf]{aastex}

2: %\documentclass[preprint2]{aastex}

3: \documentclass[12pt,preprint]{aastex}

4: 

5: %\usepackage{amssymb,latexsym,graphics,eufrak}

6: 

7: \def\re{R_{e}}

8: \def\a0{a_0}

9: \def\msun{M_{\odot}}

10: \def\kms{{\rm km~ s}^{-1}~}

11: \def\cmst{{\rm cm~ s}^{-2}~}

12: \def\vinf{V_{\infty}}

13: \def\Halpha{H$_{\alpha}~$}

14: \def\deg{^o}

15: \def\kpc{{~\rm kpc~}}

16: \def\mpc{{~\rm Mpc~}}

17: \def\mol{$M/L~$}

18: \def\av#1{\langle #1\rangle}

19: \def\cmss{cm s$^{-2}$}

20: \def\deg{^o}

21: %______________________________________________________________________________

22: 

23: \begin{document}

24: 

25: 

26: \title{{\bf MOND and the mass discrepancies in tidal dwarf galaxies}}

27: \author{Mordehai Milgrom }

28: \affil{ Center for Astrophysics, Weizmann Institute}

29: 

30: \begin{abstract}

31: I consider in light of MOND the three debris galaxies discussed

32: recently by Bournaud et al.. These exhibit mass discrepancies of a

33: factor of a few within several scale lengths of the visible galaxy,

34: which, arguably, flies in the face of the cold dark matter paradigm.

35: I show here that the rotational velocities predicted by MOND agree

36: well with the observed velocities for each of the three galaxies,

37: with only the observed baryonic matter as the source of gravity.

38: There is thus no need to invoke a new form of baryonic,

39: yet-undetected matter that dominates the disc of spiral galaxies, as

40: advocated by Bournaud et al.. I argue on other grounds that the

41: presence of such ubiquitous disc dark matter, in addition to cold

42: dark matter, is not likely.

43: \end{abstract}

44: 

45: \keywords{dark matter galaxies: kinematics and dynamics }

46: 

47: 

48: 

49: \section{Introduction}

50: Bournaud et al. (2007) have recently reported on three tidal dwarf

51: galaxies that apparently formed in the debris of the collision of

52: two galaxies: possibly of NGC 5291 with another galaxy. For reasons

53: explained cogently by the authors hardly any cold dark matter from

54: the parent galaxies, if it existed there in the first place, is

55: expected to be found in these debris. If this is indeed so, the cold

56: dark matter paradigm (CDM) predicts no mass discrepancy in these

57: dwarfs, contrary to what has been reported by Bournaud et al.

58: (2007): they find in all three dynamical masses within several scale

59: lengths that exceed the the observed baryonic masses by a factor of

60: a few.

61: \par

62: The dark-matter (DM) paradigm and MOND (Milgrom 1983a,b, see Sanders

63: \& McGaugh 2002, and Bekenstein 2006 for reviews) differ greatly as

64: regards the origin and nature of mass discrepancies they predict in

65: galaxies. In MOND, these discrepancies in a given galaxy are

66: predicted exactly from the presently observed mass distribution, and

67: are oblivious to the exact formation process of the galaxy or the

68: ensuing history. They are predicted, and are observed, to follow

69: some well defined, strict regularities: galactic analogs of Kepler's

70: laws, such as the relation between total baryonic mass and the

71: asymptotic rotation speed (the baryonic Tully Fisher relation), or

72: the onset of the discrepancy at a fixed value of the centrifugal

73: acceleration (see e.g., Milgrom 2002 for a discussion of these

74: predictions, and McGaugh 2006 for a discussion of observational

75: tests). In contrast, in the DM paradigm the mass discrepancies are

76: ratios of total (dark matter plus baryons) to the baryonic mass, and

77: depend strongly on the particular history of the galaxy since the DM

78: and the baryons are subject to different influences. The formation

79: process itself, subsequent cannibalism, mergers, and ejection of

80: baryons by cataclysmic events, such as supernovae, all greatly

81: affect the resulting mass discrepancies. This is why I believe that

82: the CDM paradigm is inherently incapable of ever predicting rotation

83: curves of individual galaxies in the way that MOND does: for most

84: galaxies we simply cannot know the crucial elements of evolution a

85: given galaxy underwent.

86: \par

87: The three reported dwarfs, and possibly others like them, are an

88: exception: if they indeed formed as described by Bournaud et al.

89: then their mass discrepancies in CDM can be predicted with some

90: certainty because the collision that led to their formation erased

91: the imprints of earlier history. According to the simulations of

92: Bournaud et al., whatever the DM halo of the parent galaxies was

93: like, as long as it was spheroidal--as predicted by CDM--the debris

94: galaxies would have hardly any DM in them, and should exhibit

95: practically no mass discrepancies. This is also in line with earlier

96: simulations referenced in the online appendices of Bournaud et al.,

97: and is easy to understand qualitatively. In contrast, all three

98: dwarfs are predicted by MOND to show appreciable mass discrepancies

99: since they are measured to have low accelerations that are rather

100: deep in the MOND regime. I show below that the MOND predictions are

101: indeed born out by the observations of Bournaud et al.. Gentile et

102: al. (2007a) have recently performed a more detailed analysis, and

103: reach the same conclusions as regards the performance of MOND.

104: \par

105: Bournaud et al. consider it a more likely explanation of the mass

106: discrepancy in the dwarfs, that they actually do contain large

107: quantities of yet-undetected matter. They advocate that at least one

108: of the galaxies partaking in the collision that begot the debris had

109: large quantities of DM in their discs--with a mass typically a few

110: times that of the visible baryons. Since this DM cannot be the

111: putative cold dark matter, which form spheroidal halos and does not

112: settle into galactic discs, Bournaud et al. opt for cold molecular

113: hydrogen, which has been considered earlier as the DM in galaxies

114: and clusters (e.g., Pfenniger Combes \& Martinet 1994).

115: 

116: In section 2 I describe the MOND results and compare them with the

117: observations. In section 3 I discuss the results and contest the

118: hypothesis of large quantities of disc dark matter in galaxies.

119: 

120: 

121: 

122: 

123: \section{MOND rotation curves}

124: 

125: To calculate the predicted MOND rotation velocities, $V$, I use the

126: Newtonian velocities, $V_N$, read from Fig. 1 of Bournaud et al.

127: (2007) in the MOND relation (Milgrom 1983b)

128: 

129: $$\mu(V^2/r\a0)V^2/r=V^2_N/r. $$

130: 

131: Here $\mu(x)$ is the extrapolating function, which I take here to be

132: $\mu(x)=x(1+x^2)^{-1/2}$  and I take the acceleration constant of

133: MOND to have the value $\a0=1\times 10^{-8}$\cmss (e.g. Bottema et

134: al. 2002). Choosing another form of $\mu(x)$ will affect the

135: predictions for the larger radii only a little since the

136: accelerations there are rather smaller than $\a0$ where $\mu$ has to

137: be nearly linear for all forms. The velocities at the inner radii

138: will be affected, but as explained below these anyway depend

139: crucially on the model adopted for the baryon mass distribution.

140: 

141: Figures 1-3 show for each galaxy the Newtonian velocities calculated

142: from the distribution of the visible matter alone as modeled by

143: Bournaud et al., the MONDian speeds calculated from the above

144: equation for the same mass distribution, and the observed rotational

145: speeds. The latter are the average of the approaching and receding

146: velocities as given in Bournaud et al.. (The differences between the

147: two sides are much smaller than the errors.) For clarity's sake I

148: have not marked the error bars for the MONDian speeds since they

149: anyhow all fall within the error bars of the measured values.

150: \par

151: In the above procedure I have adopted all the system parameters (the

152: distance,  the inclinations of the galaxies, the assumed \mol values

153: for the stellar contributions, etc.) as taken by Bournaud et al.,

154: and no attempt was made to improve the agreement by best fitting for

155: these. The baryonic mass of these galaxies are dominated by gas so

156: the exact \mol value is rather immaterial. I have also not corrected

157: for asymmetric drift (i.e., those due to velocity dispersions),

158: which is expected to be rather small for these galaxies (certainly

159: much smaller than the quoted errors). Similarly, I have ignored

160: possible corrections due to the external-field-effect (EFE) in MOND

161: (Milgrom 1983a, Brada \& Milgrom 2000a,b, Angus \& McGaugh 2007, and

162: Wu et al. 2007 ). I estimate that it is rather unimportant here

163: (i.e., the acceleration field of NGC 5291 itself, and of other

164: masses, at the dwarfs positions is smaller than the internal

165: accelerations in the dwarfs themselves). Gentile et al. (2007a)

166: study this issue in more detail. They find that indeed the impact of

167: the EFE is marginal inside the last measured point. My own estimate

168: of the EFE due to NGC 5291 is even smaller than theirs. Gentile et

169: al. calculate the far field of NGC 5291 by assuming that the

170: (deprojected) HI line width ($\Delta V_{20}$) given in Malphrus et

171: al. (1997) represents twice the asymptotic rotational speed of that

172: galaxy. This corresponds to a baryonic mass of

173: $3.2\times10^{11}\msun$, which, the galaxy being rather devoid of

174: gas, corresponds to $M/L_B\approx 16 (M/L_B)_{\odot}$; this is much

175: too high ($\Delta V_{20}/2$ could easily overestimate the asymptotic

176: rotational speed for various reasons; for example, since the galaxy

177: is rather Newtonian in the inner parts, the maximum rotation speed

178: can be quite higher than the asymptotic one). I started from the

179: luminosity of $L_B\approx 2\times 10^{10}L_\odot$ and assumed

180: $M/L_B\approx 5 (M/L_B)_{\odot}$, appropriate for the color of the

181: galaxy ($B-V\approx 1$). The same mass is gotten with $M/L_K= 1

182: (M/L_K)_{\odot}$, and it corresponds to a galaxy mass 3.2 times

183: smaller than that effectively used by Gentile et al.. Also, whereas

184: they used the projected distances from NGC 5291 to the dwarfs,

185: taking a mean of $65\kpc$ as the actual distances, I used 3-D

186: distances using the position of the galaxy with respect to the ring

187: from the model of Bournaud et al.: $114\kpc$ for N, $118\kpc$ for S,

188: and $140 \kpc$ for SW (F. Bournaud 2007, private communication). All

189: together my estimates of the field of the galaxy at the position of

190: the dwarfs are 3.2--3.8 times smaller than that of Gentile et al.,

191: rendering the effect rather negligible within the presently observed

192: dwarfs.

193: 

194: Bournaud et al. argue convincingly for inclinations around

195: $i=45\deg$ for all three dwarfs stating that they should be almost

196: aligned with the ring they are embedded in.  A large part of the

197: indicated errors on the observed velocities reflect the uncertainty

198: in $i$. This contribution to the errors should than be viewed as an

199: uncertainty in the normalization of the velocity curve, not as

200: errors on individual points.  I show in Figure 1 that for one case

201: (NGC 5291N), increasing the inclination to $55\deg$ indeed improves

202: greatly the agreement bringing the MOND velocities into practical

203: coincidence with the observed ones.

204: \par

205: Gentile et al. (2007a) suggest that the inclinations of the dwarfs

206: may differ from the model inclinations of $45\deg$ because the EFE

207: induces some precession of the disc. They thus also performed MOND

208: fits with the inclinations left free, and indeed the best fit values

209: differ somewhat from $45\deg$. This is also shown in my Figure 1

210: demonstrating that the best value for NGC 5291N is nearer $55\deg$.

211: This point might require further checking; but, as explained above,

212: my arguably more realistic estimate of the external field effect is

213: rather smaller than theirs and would give an estimate of the

214: precession period that is much longer than the time since the dwarfs

215: formed, so I do not expect this to be an important effect.

216: 

217: \begin{center}

218: \begin{figure}

219: \begin{tabular}{rl}

220: \tabularnewline

221: \includegraphics[width=0.5\columnwidth]{f1.eps} &

222: \includegraphics[width=0.5\columnwidth]{f2.eps} \\

223: \includegraphics[width=0.5\columnwidth]{f3.eps}\\

224: %\tabularnewline \R{{\footnotesize }}\tabularnewline

225: \end{tabular}\par

226: \caption{The rotation curves for the three dwarfs: NGC 5291N (upper

227: left), and 5291S (upper right), 5291SW (lower left). The measured

228: velocities, assuming the nominal inclination of $45\deg$, are marked

229: by stars and are shown with their error bars. The calculated

230: Newtonian velocities are marked by inverted triangles. The predicted

231: MOND velocities are marked by squares. The error bars for the last

232: two are omitted for clarity's sake. I also show, for NGC 5291N only,

233: as diamonds, the measured velocities for an assumed inclination of

234: $i=55\deg$.}

235: \end{figure}

236: \end{center}

237: 

238: \section{Discussion}

239: We see that the MOND predictions are in very good agreement with

240: the measured speeds. It should be noted that the baryonic

241: Newtonian curves given in Bournaud et al. are not based on an

242: actual measurement of the observed baryonic distribution. It is

243: based on taking the total amount of observed baryon and assuming a

244: model for their spatial distribution. The exact values of

245: Newtonian velocities, and hence the MOND values I deduce from them

246: for the inner radii depend on the exact distribution assumed.

247: However, at larger radii, which are already quite beyond the

248: baryonic mass concentration, the velocity values are rather

249: independent of the assumed distribution of baryons. And, after

250: all, this is where the impact of the comparison with both CDM and

251: MOND predictions comes from.

252: \par

253: The uncertainties in the analysis of the individual dwarfs are

254: relatively large. This can be traced mainly to the fact that they

255: are much farther than the dwarf spirals of similar properties that

256: have been analyzed before (e.g.,  de Blok \& McGaugh 1998, Gentile

257: et al. 2007b, Milgrom and Sanders 2007). However, while

258: individually, the data is not up to the highest standards, all

259: three dwarfs speak in the same voice: all showing a discrepancy of

260: similar magnitude developing at larger radii, with the observed

261: velocities flattening off in just the way and magnitude predicted

262: by MOND. Thus, the collective conclusion is stronger than the

263: individual ones separately. In addition, the particular importance

264: of these systems lies in their unique potential for

265: differentiating between CDM and MOND.

266: 

267: 

268: \par

269: Bournaud et al. propose another explanation of the mass discrepancy

270: in the three dwarfs: one of the colliding galaxies, which

271: contributed the gas to the debris, also harbored large quantities of

272: some form of DM (not CDM) in its disc, several times more massive

273: than the visible baryons. This DM could then have found its way into

274: the tidal dwarfs, giving rise to the observed mass discrepancy. The

275: candidate they advocate is cold, difficult to detect H$_2$. Unless

276: we want to assume that the presence of such molecular DM in the disc

277: of the parent galaxy is a rare occurrance, this DM has to be

278: ubiquitous in disc galaxies, as indeed is proposed by Bournauad et

279: al.. They note, however, that this would require a very large

280: conversion factor from the observed CO to H$_2$: an order of

281: magnitude larger than what is known for galaxies in general (see

282: also the caveats listed by Elmergreen 2007).

283: \par

284: I feel that the presence of large quantities of a new component of

285: DM in the disc of spiral galaxies, in addition to the dominant CDM,

286: is unlikely on additional grounds. It is known that the mass and the

287: distribution of the DM in disc galaxies are strongly correlated with

288: those of observed baryons through the relations predicted by MOND.

289: For example, the total mass of the latter, $M_{vis}$, is strongly

290: correlated with the asymptotic rotational speed, $V_{\infty}$ (which

291: is determined mostly by the DM) via the MOND relation (Milgrom

292: 1983b)

293: $$V^4_{\infty}=\a0 GM_{vis}.$$ This MOND relation (aka the baryonic

294: Tully-fisher relation) is found to hold over some 5 orders of

295: magnitude in galactic mass (McGaugh 2005,2006). Another correlation

296: is the onset of the mass discrepancy (equality of the contributions

297: of visible and dark matter) at a fixed value, $\a0$, of the

298: centrifugal acceleration (see McGaugh 2006 for a recent test of

299: this). But the mother of them all is the fact that the visible

300: matter distribution determines the full rotation curve of a galaxy,

301: which in the dark matter paradigm is determined by both components

302: and is dominated by DM in the outer parts. These are all

303: observational facts whatever the interpretation of MOND is. In the

304: context of CDM these correlations require various independent

305: conspiracies between the baryons and the CDM, conspiracies whose

306: origin remains a mystery (see Milgrom 2002 for a more extensive list

307: of these conspiracies and an explanation of why they are

308: independent).

309: \par

310: As emphasized many times before, these predicted MOND relations as

311: traditionally formulated are exactly valid only for completely

312: isolated systems. In the presence of an external field these are

313: modified in a manner that is also predicted by MOND (provided the

314: external field is known). However, for many systems the external

315: field effect enters importantly only at rather large radii and

316: leaves a large range of radii for which these predictions are valid

317: with high accuracy (for the Milkey Way it is expected to be

318: important only beyond a few hundred kiloparsecs). The effects of

319: external fields are deemed quite unimportant, as far as I know, for

320: all rotation curve analyses published to date, and hence are also

321: unimportant, for example, in the results of McGaugh 2006 quoted

322: above. Some exceptions concern galaxies in the cores of galaxy

323: clusters, where the external (cluster) fields are of order the of

324: $\a0$, and low acceleration systems (such as dwarf spheroidals or

325: diffuse globular clusters) in the field of galaxies.

326: 

327: 

328: \par

329: If one now adds another epicycle to the DM paradigm in the form of a

330: dominant, baryonic disc component, the observed correlations, by

331: which only the sub-dominant, visible-baryons component determines

332: everything, would require an even more involved, three-headed

333: conspiracy. This is not an argument that definitely excludes the

334: molecular-DM-plus-CDM hypothesis, but it does diminish its

335: likelihood.

336:  \par

337:   The above argument poses a difficulty for the double

338: DM hypothesis even before we consider the tidal dwarfs themselves.

339: In addition, the present rotation curve results show that the three

340: dwarfs also satisfy the above correlations. This adds another

341: dimension to the above argument: if we accept that the dwarfs formed

342: in a very different way from that of most other galaxies, and that

343: their matter component mixture is very different, why should they

344: still satisfy the same relations? In the dwarfs' case these would be

345: relations between the visible baryons and the molecular DM, while in

346: general they would be relations between the visible baryons and the

347: combined molecular-plus-cold DM. Of course, this latter part of the

348: argument assumes the robustness of the results and interpretation of

349: Bournaud et al., including their deduced inclinations. It is also

350: based, at the moment, on only the three galaxies discussed here. The

351: argument would clearly benefit from further substantiation, and the

352: examination of more tidal dwarfs. Until then it remains a tentative

353: difficulty for the double DM hypothesis.

354: 

355: 

356: \acknowledgements I am grateful to Rainer Plaga for pointing out to

357: me the potential in the results of Bournaud et al. to discriminate

358: between MOND and CDM. I also appreciate comments from Frederic

359: Bournaud and from the referee. The research was supported by a

360: center of excellence grant from the Israel Science Foundation.

361: 

362: 

363: 

364: \begin{thebibliography}{}

365: 

366: \bibitem[Agnus \& McGaugh 2007]{agnus07}  Angus, G.W. \& McGaugh, S.S. 2007, arXiv:0704.0381

367: \bibitem[Bekenstein 2006]{bek06} Bekenstein, J.D. 2006, Contemporary Physics 47, 387

368: \bibitem[Bottema et al. 2002]{bottema02}Bottema, R., Pesta\~{n}a, J. L. G., Rothberg, B., Sanders, R. H. 2002, AA, 393, 453

369: \bibitem[Bournaud et al. 2007]{bournaud} Bournaud, F., Duc, P.-A.,

370: Brinks,E.,  Boquien, M., Amram, M., Lisenfeld, U., Koribalski, B.S.,

371: Walter, F., Charmandaris, V. 2007, Science, 316, 1166

372: \bibitem[Brada \& Milgrom 2000a]{bm00a}Brada, R. \& Milgrom, M. 2000a ApJ, 541, 556

373: \bibitem[Brada \& Milgrom 2000b]{bm00b}Brada, R. \& Milgrom, M. 2000b ApJ, 531, L21

374: \bibitem[de Blok \& McGaugh 1998]{dbmg98}de Blok, E. \& McGaugh, S.S. 1998, ApJ, 508, 132

375: \bibitem[Elmergreen 2007]{elmer07}Elmergreen, B.G. 2007, Science (perspectives), 316,

376: 1132

377: \bibitem[Gentile et al. 2007a]{gentile07a} Gentile, G., Famaey, B.,

378: Combes, F., Kroupa, P., Zhao, H.S., \& Tiret, O. 2007a, AA, in

379: press, arXiv:0706.1976

380: \bibitem[Gentile et al. 2007b]{gentile07b} Gentile, G., Salucci, P.,  Klein, U., and

381: Granato, G. L. 2007b MNRAS 375, 199

382: \bibitem[Malphrus et al. 1997]{malphrus97} Malphrus, B.K., Simpson,

383: C.E., Gotesman, S.T., \& Hawarden, T.G. 1997, AJ 114, 1427

384: \bibitem[McGaugh 2005]{mcg05}  McGaugh, S.S. 2005, ApJ, 632, 859

385: \bibitem[McGaugh 2006]{mcg06}  McGaugh, S.S. 2006,

386: astro-ph/0606351

387: \bibitem[Milgrom 1983a]{mil83a} Milgrom, M. 1983a, ApJ, 270, 365

388: \bibitem[Milgrom 1983b]{mil83b} Milgrom, M. 1983b, ApJ, 270, 371

389: \bibitem[Milgrom 2002]{mil02} Milgrom, M. 2002, New Astron.Rev. 46 741

390: \bibitem[Milgrom \& Sanders 2007]{ms07}Milgrom, M. \& Sanders, R.H. 2007 ApJ, 658, L17

391: \bibitem[Pfenniger Combes \& Martinet 1994]{pfenniger94}Pfenniger,

392: D., Combes, F., \& Martinet, L. 1994, AA, 285, 79

393: \bibitem[Sanders \& McGaugh 2002]{sm02}Sanders, R.H. \& McGaugh, S.S. 2002, ARA\&A, 40, 263

394: \bibitem[Wu et al. 2007]{wu07} Wu, X., Zhao, H.S., Famaey, B.,

395: Gentile, G., Tiret, O., Combes, F., Angus, G.W., \& Robin, A.C.

396: 2007, arXiv:0706.3703v2

397: 

398: 

399: 

400: \end{thebibliography}

401: 

402: \end{document}

403: