in the : 60 Years Later

J aan Einasto Astrophysical , EE-2444,

1 Introduction

The traditional character of has been changed by the discoveries of dark matter in the Universe and of the filamentary structure of the Universe with euormous empty regions devoid of any kind of visible matter between filaments. In this year we celebrate the 60 anniversary of the disc.overy of the dark matter. 1 devote this review to some aspects of the dark ma.tter problem. The term dark matter is used here for any form of matter whose exi­ stence ca.n be detected only through its gravitational effect~. The presence of dark matter, its quantity and extent can be determined from the dif­ ference between the dynamical mass estimates in stellar systems and the directly counted masses of visible objects in them. If a difference between dynamical masses and directly counted masses is detected we can speak on the mass paradox or on the "missing mass". Mass determinations in smaU astronomie.al systems as the solar system, binary stars and sma.11 stellar systems show good mutual consistency bet­ ween results obtained by different methods and there is no room left for dark matter. A mass paradox is ·found in gala.ctic disks, around galaxies, in gro­ ups, clusters and superclusters of galaxies. The mass paradox is solved by assuming the existence of dark ma.tter in respective systems. Dark mat­ ter exists probably a.lso in large voids between superclusters. Dynamically dark matter is dominant on a.11 se.ales larger than the sca.le-length of ordina-

17 19 18 Jaan Einasto Dark Mat ter in the Uni verse: 60 Years La ter

ry galactic populations. Thus the formation of galaxies and the structure If velocitie and orbital planes of test particles are distributed rando~y, of the Universe is essentially determined by the properties of dark matter. then we have another simple case. From measurements of Dopple~ shifts This makes the problem of the amount and nature of dark matter to the we can determine the velocity dispersion in one coordinate, Un direct~d central one in the whole cosmology. towards tlLe observer. Now instead of the velocity, V, we have the radial In this review I consider first shortly the methods of mass determina­ velocity dispersion, u, in this case f3 = 1/3, and we get tions in various stellar systems. Then a short history of the study of the 2 - 1/3GM(R) (3) dark matter is given. Current topical problems of the dark matter are ur - R . considered next: What is the amount of matter associated with galaxies? How much matter can be located in voids? What is the mean density of This case can be appliod for spheri al lusters of stars or galaxies. U sing t h.e matter in the Universe? What is the nature of dark matter? associated luminosity obtained by photometry one can deduce the mass­ to-luminosity ratio, M / L, which characterizes the physical n.ature of the 2 Mass Determinations of Stellar Systems system. The velocities ca.n be measUTed cLirectly from redsrufts, ~ut the ·t· ·lZ R aii

usually measured in solar units. . _ 2 where M(R) is the mass inside the orbit of test particles of radius R, Gis In cosm logy it is customa.ry to calculate the quant1ty n =87rGe/(3JI_ ), the gravitational constant, and f3 is a dimensionless constant of the order the mean mass density, ë, roeasured in units of the critical closure _

was asked by Oort (1932). New determinations were made by Hill (1960) tluough in the dark matter problem came when Tartu and Princeton gro­ and 0.ort .(1960). They noticed that the dynamical density of mass in the ups demonstrated that dark coronas or haloes around giant galaxies have galactic d1sk exceeds the density of directly observed matter, the discre­ an extent of several hundred kiloparsecs and total masses of galaxies exceed pancy was only·by two times. Thls invisible component must form a disk · conventional masses tenfold (Einasto, Kaasik Saar 1974, Ostriker, Peebles, otherwise it distorts the whole rotation curve of the Galaxy. Presentl; Yahil 197 4 ). This discrepancy can be explained by the presence of a mas­ the problem of the presence of such "local dark matter" is not finally sol­ sive invisible corona around the galaxy. These results were confirmed by ved: observation~ determinations are rather uncertain, and the discussion, a systematic survey of rota.tian curves of galaxies by Rubin et al. (1980, whether there eXJsts local dark matter in the Galaxy or not, continue. 1'982). Reviews of the dark matter problem in various systems of galaxies . The ~ctual discovery of the "missing mass" concerns clusters of gala­ are written by Fa.ber and Gallagher (1979) and Trimble (1987). XJes: Zwicky (1933), basing on a virial analysis of 7 galaxies in the Coma There may exist a general srnoothly distributed dark matter backgro­ dus ter, argued for a discrepancy by a factor of rv 400 between the viri­ und too. Studies of the large-scale structure of the Universe demonstrate al M/ L in Coma and that of nearby galaxies. This was an overestimate that galaxies and clusters of galaxies are located in :filaments which form because it was based on a wrong value of the Hubble constant, H = 500 a connected 3-dirnensional network (a review is given by Oort 1983). The km~s/Mpc. According to modern data the mass-to-luminosity ratio in Co­ space between filaments is void of any visible matter. Galaxy systems :fi.11 ma 1s by a factor of~ 100 larger than in typical galaxies. This measures only a few per cent of the total volume of the space (Einasto, Jôeveer, Sa­ the amount of the "missing mass" in clusters of galaxies. ar 1980) a.nd diameter of voids reach tens megaparsecs (Jôeveer, Einasto, A similar discrepancy has been discovered in groups of galaxies. Kahn Tago 1978, Kirshner et al. 1981). A complete evacuation of voids of such and Woltjer (1959) noticed that the mass of the Local Group of galaxies, large diameter is impossible since at the epoch of recombination the den­ as determined from the motion of our Galaxy in respect to the Andromeda sity distribution was still almost homogeneous, and gravitation, the only 12 galaxy, is of the order of 10 M0 whereas the masses of bath galaxies are· 11 force which eau evacuate voids, works slowly. We corne to the conclusion of the order of 10 M0. We see a discrepancy of the order of ten. that some dark matter should exist in voids. The next stage of the dark matter story was connected with individual Summarizing the dynamica.l evidence we can say that there are actually galaxies. The mass distribution in galaxies can be determined directly from three dark matter problems: (1) dark matter in galactic disks; (2) dark their rotation curves and indirectly from the light distribution in visible ma.tter associated with galaxies and clusters of galaxies; (3) a smoothly populations using photometric data. In late 60s both data were available distributed background of dark matter. for a. few nea.rby galaxies and a comparison of photometric and dynamical modela was possible. This comparison demonstrated the presence of mass paradox in spiral galaxies. Einasto (1969, 1972), Sizikov (1969), Freeman 4 Mean Density of Matter Associated with Ga­ (19~0),, Rubin and Ford (1970), Rogstad and Shostak (1972) independen­ laxies tly md1cated tha.t from photometric data one should expect a Keplerian falloff of the circular velocity on the periphery of spiral galaxies, whereas Most galaxies are loca.ted in systems: groups, clouds and clusters. ThesE\ dynamical data indicate only a marginal if any decrease of the rotation systems are distributed along some sort of connected network which leaves velocity. most of the Universe empty. It is surprising that neither the Zwicky's discovery nor the Kahn and Gramann (1990) estimated the density of ma.tter associated with sy­ Woltjer's paper and accumulating evidence on the presence of the mass stems of galaxies. At present, no detailed dynamical information on ail paradox in galaxies have been given attention they deserved. A real break- galaxy systems in a large volume is available. However, in a large region we can estima.te the mean luminosity density. 22 Jaan Einasto Dark Matter in the Universe: 60 Years Later 23

The mean denshy of matter associated with systems of galaxies in the The main contribution to ne cornes from groups. If a group is bound, given volume V is deterrnined as bµt not in virial equilibrium, the application of the virial theorem for the estimate of its mass is not correct. The true median M / L of groups may (4) .. be ' greater than the virial estimates. In this case ne increases ( Giuricin et al ..1988). Here (M/ L); and f; are the mass-to-lurninosity ratio and frequency of If we take into account the uncertainty in bath the mean luminosity sys_terns of different types i, respectively, and eL is the mean luminosity density and the mass-to-luminosity ratio, and possible systematic error in density in the given volume. M/ Lof groups, we get an estimate ne= 0.15 ± 0.05. ~vailable estimates of the luminosity density by Jôeveer et al. (1978), Davis and Huchra (1982), Efstathiou et al. (1988) yield a mean value about 8 3 5 Mean Density of Matter in Voids and in the {!L ~ 1.5 · 10 h Mpc- . Luminosity density is given in the Zwicky B(O) system. Universe Adopting the value given above as the characteristic mean luminosity The mass budget of the Universe has three basic constituents: the lumi­ density in the region containing one or two superclusters, the clustered nous mat ter (galaxies, intergalactic gas ); clustered matter associated with matter density parameter ne is galaxies and systems of galaxies; and non-clustered matter in voids. Mass-to-luminosity ratios of luminous galactic populations are in the 4 ne= eM-87rG = (6 ± 2) · 10- "' (M)- f· h-1 (5). 3JI 2 L.J L . i range 1-10, and we can adopt a median value 3. If we add to this amount i ' the intergalactic gas we get the total mean density of the luminous mat ter,

The mass-to-luminosity ratio M/L is expressed in solar units (M/L) . n1um l'V 0 0.01. We eau distinguish tluee types of systems: rich clusters, groups and The clustered matter, associated with galaxies and systems of galaxies field galaxies. The median mass-to-luminosity ratio for rich clusters has gives, has, according to data discussed in previous Section, .a contribution been estimated as M/ L ~ 500 ± lOOh by Geller (1988). Mean masses and ne ~ 0.15. This contribution ïs divided into two parts, given by luminous mass-to-luminosity ratios of groups of galaxies can be determined using and dark matter, in a ratio 1 to 15. published catalogues of groups (Geller and Huchra 1983, Vennik 1984, Voids and superclusters are made of initial negative and positive density Tully 1987). Such determinations have been performed by several authors fluctuations, respectively. In a random Gaussian fluctuation field positive (Huchra and Geller 1982, Vennik 1986), and yield a mean M/L = 190 h. and negative fluctuations are interchangeable; thus the amount of mat­ ln the case of field galaxies we can use rotation curves. As demonstrated ter in both types of fluctuations is initially equal. During the subsequent by Rubin et al. ( 1985 ), isolated galaxies also have fiat rotation curves and evolution matter fl.ows from low-density regions toward high-density re­ thus are surrounded by dark coronas. Conservative extrapolation of mass gions. The rate of fl.ow of matter from low density regions can be found fonctions M(R) yields a mass-to-luminosity ratio of the order of 100 h. numerically. From a full galaxy catalogue used in a group and cluster search, about Numerical simulations by Einasto et al. (1994) show that the fraction of 80% of galaxies belong to groups, and about 5 % belong to clusters, the pa.rticles in low-density regions ( voids) is presently Fv ~ 0.15, respectively rest are relatively isolated galaxies. Using these data we can adopt the the total amount of matter in voids 'in units of the critical density nv ~ following frequencies for clusters, groups and field galaxies: 0.05; 0.80; 0.03. The density value is fairly robust, for a wide range of models results 0.15. If we adopt the mean mass-to-light ratios cited above, we get for the are almost identical. mean mass-to-light ratio M/ L = l90h and ne= 0.11. 24 Jaan Einasto Da.rk Ma.tter in tbe Universe: 60 Years La.ter 25

The total amount of matter in the Universe, including both clustered There exist no known physical processes which can segregate the smoothly and non-clustered matter, is no = ne+ nv ::;;:j 0.2. distribute

6 Nature of Dark Matter [1] Davis, M., Huchra, J. 1982, Ap. J. 254, 425.

The local dark matter (if it exists) is without any doubt baryonic. Its [2] Efstathiou, G., Ellis, R.S., Peterson, B.A. 1988, Mon. Not. R. Astr. contribution to the total mass budget of the Universe is rather small. Soc. 232, 431. Nucleosynthesis studies suggest that the total density of baryonic mat­ [3] Einasto, J. 1969, Astrofiz. 5, 137. ter must be nbar ~ 0.05 - 0.1. This estimate is not very different from the presently estimated total amount of matter, n0 ~ 0.2. Thus dark mat­ [4] Einasto, J. 1972, Tartu Astr. Obs. Teated 40, 3.; Proc·. First Europ. ter in coronas of galaxies and clusters may consist of some non-dissipative Astr. Meet., ed. L. N. Mavrides, Springer, 2, 291. form of baryonic abjects. The real problem here is how to transfer or­ dinary primeval gas in a very early stage and with high effi.ciency to the [5] Einasto, J., Jôeveer, M., Saar, E. 1980, Mon. Not. R. Astr. Soc. 193, dark form and then stop the process abruptly to explain the gap between 353. the physical and dynamical properties of dark matter and ordinary stellar [6] Einasto, J., .Kaasik, A., Saar, E. 1974, Nature 2.50, 309. populations. The dark matter may as well be non-baryonic, except that massive [7] Einasto, J., Saar, E., Einasto, M., Freudling, W., Gramann, M. 1994, neutrinos (hot dark matter) cannot be confined to small coronas around Ap. J. (in press); ESO Preprint no. 971. galaxies because of phase-space density constraints (Tremaine and Gunn 1979). Thus non-baryonic dark coronas of galaxies must be made of some [8] Faber, S.M. and Galla.gher, J.S. 1979, Ann. Rev. Astron. Astrophys. sort of cold dark matter which can form small coronas too. 17,1:35. Dark mat ter in voids must be of the same type as in the clustered com­ [9] Freeman, K. C. 1970, Ap. J. 160, 811. ponent around galaxies due to the continuons exchange of matter between voids and galaxy systems. A certain fraction of it can be baryonic. This [10] Geller, M. 1988, in Large Scale Structure and Motions in the Universe, fraction must be identical in the clustered and non-clustered component. eds. M. Mezzetti et. al. , Kluwer, Dordrecht, p. 3. 26 Jaan Einasto Dark Matter in tl1e U11iverse: 60 Yeàrs Later 27

[11) Geller, M., Huchra, J. 1983, Ap. J. Suppl. 52, 61. [30] Tremaine, S. and Gunn, J. E. 1979, Phys. Rev. Lett. 42, 407.

[12] Giurîcin, G., Gondola, P., Mardîrossian, F., Mezettî, M., Ramella, M. [3~] Trimble V. 1987, Ann. Rev. Astron. Astrophys 25, 425. 1988, Astr. Astrophys. 199, 85. [32] Tully, R.B. 1987, Ap. J. 321, 280. [13] Gramann, M. 1990, Mon. Not. R. Astr. Soc. 244, 214. [33] Vennik, J. 1984, Tartu Astr. Obs. Teated No. 73. (14] Hill, E. R. 1960, Bull. Astr. Inst. Neth. 15, 1. [34] Vennik, J. 1986, Astron. Nacllf. 307, 157. [15] Huchra, J., Geller, M. 1982, Ap. J. 257, 423. [35] Zwicky, F. 1933, Helv. Phys. Acta 6, 110. [16] Jôeveer, M., Einasto, .J., Tago, E. 1978, Mon. Not. R. Astr. Soc. 185, 35.

[17] Kahn, F. D., Woltjer, L. 1959, Ap. J. 130, 705.

[18] Kîrshner, R. F., Oemler, A., Schechter, P. L., Schectman, S. A. 1981, Ap. J. 248, 157.

[19] Oort, J. II. 1932, Bull Astr. Inst. Netli. 6, 249.

[20) Oort, J, H. 1960, Bu11 Astr. Inst. Neth. 15, 45.

[21] Oort, J. H. 1983, Ann. Rev. Astron. Ap. 21, 373.

[22] Ôpik, E. 1915, Bull. de la Soc. Astr. de Russie, 21, 150.

[23] Ostriker, J. P., Peebles, P. E., Yahil, A. 1974, Ap. J. 193, Ll.

[24] Rogstad, D. H., Shostak, G. S. 1972, Ap. J. 176, 315.

[25] Rubin, V C., Burstein, D., :Ford, W. K., Thonnard, N. 1985, Ap. J. 289, 81.

[26] Rubin, V. C., Ford, W. K. 1970, Ap. J. 159, 379.

[27] Ruhin, V. C., Ford, W. K., Thonnard, N. 1980, Ap. J. 238, 471.

[28] Rubin, V. C., Ford, W. K., Thonnard, N., Burstein, D. 1982, Ap. J. 261, 439.

[29] Sizikov, V. S. 1969, Astrofiz. 5, 317.

3*