Low Density Structures in the Local Universe. I. Diffuse Agglomerates of Galaxies
I. D. Karachentsev, V. E. Karachentseva, O. V. Melnyk, A. A. Elyiv, D. I. Makarov
aa r X i v : . [ a s t r o - ph . C O ] O c t Low Density Structures in the Local Universe.I. Diffuse Agglomerates of Galaxies
I. D. Karachentsev, V. E. Karachentseva, O. V. Melnyk,
3, 4
A. A. Elyiv,
2, 4 and D. I. Makarov Special Astrophysical Observatory of the Russian AS, Nizhnij Arkhyz 369167, Russia Main Astronomical Observatory, National Academy of Sciences, Kiev, 03680 Ukraine Astronomical Observatory, Taras Shevchenko National University of Kiev, 04053 Ukraine Institut d’Astrophysique et de Geophysique,Universit´e de Li`ege, Li`ege, B5C B4000 Belgium (Received August 30, 2012; Revised September 7, 2012)This paper is the first of a series considering the properties of distribution of nearby galaxiesin the low density regions. Among 7596 galaxies with radial velocities V LG < km/s,absolute magnitudes M K < − . m , and Galactic latitudes | b | > ◦ there are 3168 fieldgalaxies (i.e. 42%) that do not belong to pairs, groups or clusters in the Local universe.Applying to this sample the percolation method with a radius of r = 2 . Mpc, we found226 diffuse agglomerates with n ≥ number of members. The structures of eight mostpopulated objects among them ( n ≥ ) are discussed. These non-virialized agglomeratesare characterized by a median dispersion of radial velocities of about km/s, the linearsize of around Mpc, integral K -band luminosity of × L ⊙ , and a formal virial-mass-to-luminosity ratio of about M ⊙ /L ⊙ . The mean density contrast for the consideredagglomerates is only h ∆ n/n i ∼ , and their crossing time is about – Gyr.
1. INTRODUCTIONRecent photometric and spectral massive sky surveys: 2MASS [1], SDSS [2], 2MRS [3], 6dF [4]etc. demonstrate that the principal elements of the large-scale structure of the Universe are cosmicvoids, embordered by filaments and walls towards which the galaxies are concentrated. In knots,at the intersection of walls and filaments rich clusters of galaxies are emerging. The glow of hotintergalactic gas in rich clusters makes them outstanding objects in the X-ray maps of the sky.Numerical simulations of the formation and evolution of the large-scale structure sustain this patternquite well [5–8]. In the modern epoch ( z = 0 ) virialized regions of groups and clusters of galaxies,as well as the collapsing regions around them concentrate about 74% of all the galaxies, or about90% of stellar mass. However, these dynamically “advanced” regions occupy only about 5% of thetotal volume [9]. The remaining 95% of the volume are occupied by about a quarter of all galaxies(or 10% of stellar mass) which are involved in the infinite cosmic expansion. % N K, mag %
100 100 474852505451535762697685 819097 N M K , mag Figure 1.
The distribution of the number of galaxies with radial velocities V LG < km/s by apparent( K ) and absolute ( M K ) magnitudes in the K -band. Clustered galaxies are hatched, their relative number ineach interval is shown in percentage at the top edge. At present we are witnessing a somewhat paradoxical situation: rich clusters of galaxies, theirstructure and evolution are already investigated with enough detail, while the properties of theprincipal elements of the cosmic volume (the voids, filaments and walls) are so far only known inthe most general outline. The emphasis made on the study of “tops” of the large-scale structureand the neglect of its “roots” renders the prevailing approach quite asymmetric. One reason for thisasymmetry is the paucity of data available on the individual distances of galaxies. The ExtragalacticDistance Database ( http://edd.ifa.hawaii.edu ), created by Tully et al. [10] shows that therelative number of galaxies with measured distances rapidly drops with increasing distance, makingup a small percentage as early as at D ∼ Mpc.To study the properties of the nearby part of the large-scale structure, we compiled a sample of10 500 galaxies with radial velocities of V LG < km/s relative to the centroid of the Local Group(LG), which covers the entire sky except for low Galactic latitudes | b | < ◦ . The morphologicaltypes, the data on the radial velocities and apparent magnitudes were determined or refined for allthe galaxies of this sample. Using the new clustering algorithm, which takes into account the galaxydifferences by luminosity, there were compiled the catalogs of 509 pairs [11], 168 triple systems [12]and 395 groups of galaxies [13]. In addition, a separate catalog was devoted to 520 most isolatedgalaxies in this volume [14].Since the best indicator of the stellar mass of the galaxy is its K s -band luminosity, we adopted theapparent K s -magnitudes of galaxies from the 2MASS survey [1]. In their absence, the K -magnitudes Figure 2.
The distribution of clustered (top) and non-clustered (bottom) galaxies with V LG < km/sin equatorial coordinates. The region of significant galactic absorption with A g > . m is described by a grayragged stripe. of galaxies were determined from the known B -magnitudes and the mean color indices h B − K i depending on the morphological type of the galaxy. From the original sample of 10 500 galaxies with V LG = [0 – km/s, we have eliminated the objects fainter than K = 15 . m , and dwarf galaxieswith absolute magnitudes of M K > − . m given the Hubble constant of H = 73 km/s/Mpc. Thelatter condition provides that the galaxies with luminosities brighter than the luminosity of theSmall Magellanic Cloud would be visible both nearby and at the far edge ( m − M = 33 . m ) of theconsidered volume. Having sacrificed 2906 dwarf galaxies (28% of the sample), we have considerablyrelaxed the selection effect by distance, which was imposing unequal conditions to the nearby andfar volumes. We use the sample of 7596 galaxies corrected this way for the further analysis ofelements of the large-scale structure of the Local universe at extremely low densities.2. THE FIELD GALAXIES AND CLUSTERED POPULATIONThe distribution of 7596 galaxies in our sample by the apparent ( K )and absolute ( M K ) magnitudes is presented in the left and right pan-els of Fig. 1. The maximum of the N ( K ) distribution falls on K ≃ . , from what we can conclude that in many galaxies of this volume with K = 12 . – . theline-of-sight velocities have not yet been measured. The clustered galaxies are marked by hatchingin both panels. The clustering algorithm applied assumes that the stellar mass of each galaxy isdetermined by its K -luminosity, and the total mass of its dark halo is κ = 6 times larger than thestellar mass. The criterion of inclusion of galaxies in a pair or group was based on two obviousconsiderations: 1) the total energy of a hypothetic system has to be negative, and 2) the membersof a virtual system have to be causally interrelated (their mutual separations have to be withinthe “zero-velocity sphere”, which separates a potential group from the common Hubble expansion).The latter condition is required as a complement to the former, since we do not know the total(spatial) distances and velocities of galaxies. In fact, our algorithm has an only arbitrary parameter κ , assumed to be equal to 6 regardless of the luminosity of a galaxy and its neighborhood. Notethat the global ratio of the dark matter density to the baryon density is Ω m / Ω b ≃ [15].The application of our algorithm resulted in the integration of 4428 galaxies out of 7596 intosystems, i.e. the percentage of clustered galaxies and members of the general “field” amounted to58:42. It proved to be slightly higher than for the entire original sample, 52:48, from which the dwarfobjects were not yet eliminated. These figures suggest that normal galaxies tend to get clusteredmore than their dwarf counterparts.The sky distribution of 4428 clustered and 3168 non-clustered galaxies of the Local universe ispresented in the top and bottom panels of Fig. 2 in equatorial coordinates. The region of significantgalactic absorption along the Milky Way is shown by the gray ragged stripe. As we can see, themembers of systems of different multiplicity are revealing a strong concentration towards the equatorof the Local Supercluster, centered in the Virgo cluster (12 h m +12 ◦ ). The population of non-clustered galaxies practically does not show this concentration. At the same time, the distributionof field galaxies does not look quite uniformly random. In different regions of the sky low contrast Parameters of the most populated agglomerates in the Local universe
Agglpmerate RA DEC n < V LG > σ v < r > L K M V IR M V IR /L K n(E,S0)km/s km/s Mpc L ⊙ M ⊙ Leo–Virgo . h + 4 ◦
83 +1210 158 8.7 3.4 2.4 700 9Eridanus–Columba 4.3 –36 69 +1080 273 7.6 3.0 6.2 2050 13Centaurus 13.3 –32 43 +2310 182 6.5 4.3 1.6 360 4Microscopium 21.0 –39 39 +2670 110 6.2 3.0 0.8 270 3Crater–Corvus 11.9 –17 33 +1510 180 4.8 1.7 1.7 1000 1Libra–Hydra 15.1 –20 29 +2300 153 6.2 2.3 1.6 690 2Virgo 12.4 +2 25 +2070 217 4.7 2.0 2.4 1210 1Tucana–Grus 22.5 –59 25 +3170 109 4.1 5.1 0.6 110 2 N V LG , km/s % Figure 3.
The distribution of galaxies in the Local universe by radial velocities. Clustered galaxies arehighlighted by hatching, their relative number in each velocity interval is marked in percentage at the topedge. structures are visible, the presence of which is not related to the flocky galactic extinction.Figure 3 represents the line-of-sight velocity distribution of galaxies in our sample in the LGcentroid frame. Clustered galaxies are marked by hatching. Their relative number next to theulterior boundary of the volume significantly drops due to the shortage of galaxies with measuredvelocities among the distant objects.As we know, the galaxies of early morphological types (E, S0, Sa) show a higher degree ofclustering than the late-type galaxies. The expected effect of morphological segregation is alsoevident in our samples. Figure 4 shows the distribution of galaxies of our volume by morphologicaltypes in the de Vaucouleurs scale. The clustered galaxies are marked by hatching, their percentage -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 100100200300400500600700800900 494940505054585860686772747881 N T % Figure 4.
Distribution of the number ofgalaxies by morphological type. Clusteredgalaxies are hatched, their percentage in eachtype is indicated at the top edge. N r , Mpc r =2.8 Mpc Figure 5.
Distribution of non-clusteredgalaxies by distance to the nearest neighbor. for each type is shown at the top edge. As we can see, the galaxies with developed bulges (
T < )are present among the clustered galaxies in a much greater proportion than among the galaxies ofthe field. 3. PERCOLATION AND LOW DENSITY GALAXY AGGLOMERATESTo make headway in understanding the features of the distribution of 3168 non-clustered galaxies,we tried to sort out any non-random structures among them. This can be done by different means.
10 20 30 40 50 60 70 80020406080100120140160 N n Figure 6.
The number of low densityagglomerates, identified by percolation,depending on the number of galaxies in them. N R , Mpc Figure 7.
Distribution of the agglomerateswith n ≥ number of members by the averagemutual separation of galaxies. The mostpopulated structures with n ≥ are shaded. We used the simplest method of percolation. Estimating the distances to galaxies by their radialvelocities D = V LG /H at H = 73 km/s/Mpc and neglecting their peculiar velocities, we havedetermined the spatial distance r to the nearest neighbor for each of them. The distribution ofthe number of galaxies in increments of r is shown in Fig. 5 in the logarithmic scale. Two-thirdsof the galaxies have their nearest neighbor within r = 2 . Mpc. We selected this value as thepercolation radius via the “friends of friends” (FoF) method [16].Combining the galaxies with mutual separations of r < . Mpc into the agglomerates ofdifferent populations n , we obtained the following result: the number of non-percolated, i.e. veryisolated galaxies was found to be 543. The remaining galaxies have grouped into agglomerates withthe number of members from 2 to 84. The distribution of the number of such structures by thenumber of members is shown in Fig. 6. It is easy to see that compared with the Poisson distributionthis one has a long tail, the presence of which indicates that the unification of galaxies in largeassociations is not random.Figure 7 shows the distribution of 226 agglomerations with n ≥ members, as well as 54 mostpopulated structures with n ≥ by the average mutual separation of their members. The mediansof these distributions are 2.9 Mpc and 4.2 Mpc, respectively, i.e. the linear dimensions of thesestructures are comparable to the virial radius of rich clusters of galaxies. Figure 8.
The sky distribution of galaxies, belonging to the agglomerates with n ≥ members (top panel)and galaxies, not subjected to percolation (bottom panel), in equatorial coordinates. The sky distribution of 989 galaxies belonging to the agglomerates with n ≥ members isdemonstrated in the upper panel of Fig. 8. For a comparison, the bottom panel displays a similardistribution of 543 single galaxies, not subjected to percolation ( n = 1 ). The characters of thesedistributions are strikingly different, once again suggesting that the low density regions are hostingsome non-virialized extended structures, which comprise a significant number of galaxies.The flocky nature of the distribution of galaxies belonging to the agglomerates with populationsof n ≥ is also clearly visible in Fig. 9, where we used the Cartesian equatorial coordinates. The Figure 9.
The spatial distribution of members of the low density agglomerates with n ≥ presented inthree planes of equatorial Cartesian coordinates (in Mpc). spottiness of these projected distributions is partly caused by the presence of the extinction regionin the Milky Way. However, the filamentary structure of a great many agglomerates can not becaused by the effect of galactic extinction only.4. THE MOST POPULATED GALAXY AGGLOMERATESA summary of eight diffuse agglomerates in the Local universe with the n ≥ number of galaxiesis presented in the table. The columns of the table include: (1) the names of the constellations, wherethe agglomerate is located, (2) the equatorial coordinates of the centroid, (3) the number of memberswith measured radial velocities, (4) the mean line-of-sight velocity relative to the Local Group,(5) the line-of-sight velocity dispersion, (6) the mean spatial separation between the agglomeratemembers, (7) integral K -band luminosity (i.e. the total stellar mass), (8) the formal value of virialmass, (9) the formal virial mass-to- K -luminosity ratio (or the ratio of dark-to-luminous matter),(10) the number of the agglomerate members of E and S0 morphological types. As we can see, therelative number of early-type galaxies in these structures is only about 10%.The panels of Figs. 10 and 11 show the distribution of galaxies of eight most populated agglom-erates in the projections of Cartesian equatorial coordinates (in Mpc). As follows from these figures,the shapes of agglomerates are very diverse and are generally far from being spherically symmetric.In a half of cases, they could be conditionally called filamentary or deplanate.The median line-of-sight velocity dispersion in rich agglomerates ( km/s) and median K -lumino- sity (3 . × L ⊙ ) are close to the corresponding median values for the Makarov-Karachentsev (MK) groups in the same volume of Local universe [13], but the linear dimensions ofthe agglomerates exceed the typical size of MK-groups by an order of magnitude. The median of0 Figure 10.
The structure of eight most populated low-density agglomerates in the projections of equatorialCartesian coordinates (in Mpc). The first part, continued in Fig. 11. Figure 11.
The structure of eight most populated low density agglomerates in the projections of equatorialCartesian coordinates (in Mpc). The second part, continued from Fig. 10. . × M ⊙ is comparable to the mass ofpoor clusters, whereas the median of the formal virial mass-to- K -luminosity ratio, equal to about M ⊙ /L ⊙ is by an order of magnitude higher than the corresponding value for the richest clusters.It should be emphasized that the considered agglomerates are extremely incoherent buildupswith no obvious signs of concentration of galaxies towards their geometric centers. The averagenumber density of galaxies in them is only about five times higher than the mean number density ofgalaxies in the considered volume of the Local universe. Given the scales and line-of-sight velocitydispersions specified in the table, the characteristic crossing time in these aggregates is 30–40 Gyr,what is significantly larger than the age of the Universe.5. CONCLUDING REMARKSThe considered volume of the Local universe with the diameter of about 100 Mpc is a quiterepresentative sample, including the entire Local Supercluster and the spurs of other neighboringsuperclusters. This volume comprises both groups and clusters, as well as the cosmic voids. Search-ing for the diffuse associations of galaxies in the regions of low density, we used only the kinematicdistances of galaxies, D = V LG /H , neglecting their peculiar velocities. Until fairly recently, therewas an idea that large peculiar velocities of galaxies are only found in the “hot” virial regions ofclusters, while in the field galaxies the deviations from the Hubble relation V = H D are small.According to the current data, however, the field galaxies, surrounding the Local Group are in-volved in a bulk motion towards the Virgo cluster with the velocity of about km/s, and in themotion from the center of the Local Void, caused by the expansion of the void with the velocity ofabout km/s [17]. Numerous simulations of the evolution of large-scale structure [5, 6] reveal thepresence of coherent motions of field galaxies with amplitudes of a few hundred km/s on the scaleof approximately ( – ) Mpc. The observational data on large peculiar motions of galaxies in theComa I region give indications that there exists a possible “dark attractor” with a mass of around M ⊙ at the distance of 15 Mpc from us [18].Large-scale flows of galaxies related to the motions of filaments and walls can lead to phantomphase groupings of galaxies, if only the kinematic distances are used for their clustering. Such false“phase caustics” can be easily confused with the scattered physical groupings of galaxies. Therefore,some or even most of the discussed agglomerates in the low-density regions can prove to be phantomstructures.It is obvious that for checking the verity of the existence of the diffuse agglomerates we have3discovered, the measurements of galaxy distances by the Tully-Fisher method [19] or any othertechnique, independent of the line-of-sight velocities are yet required.ACKNOWLEDGMENTSThis study was made owing to the support of the following grants:the grants of the Russian Foundation for Basic Research (project nos.11-02-90449-Ukr-f-a, 12-02-91338-NNIO), the Ukrainian State Fund for Fundamental Research(project no. F40.2/049), the Cosmomicrophysics program of the National Academy of Sciences ofthe Ukraine, and by the Ministry Education and Science of the Russian Federation (state contractno. 14.740.11.0901).
1. T. N. Jarrett, T. Chester, R. Cutri, et al., Astronom. J. , 2498 (2000).2. K. N. Abazajian, J. K. Adelman-McCarthy, M. A. Agueros, et al., Astrophys. J. Suppl. , 54 (2009).3. J. P. Huchra, L. M. Macri, K. L. Masters, et al., Astrophys. J. Suppl. , 26 (2012).4. D. H. Jones, M. A. Read, W. Saunders, et al., Monthly Notices Roy. Astronom. Soc. , 683 (2009).5. A. Klypin, Y. Hoffman, A. V. Kravtsov, and S. Gottloeber, Astrophys. J. , 19 (2003).6. W. Schaap, PhD Thesis (Groningen Univ., 2007).7. S. Shandarin, S. Habib, and K. Heitmann, arXiv:1111.2366 (2011).8. A. A. Klypin, S. Trujillo-Gomez, and J. Primack, Astrophys. J. , 102 (2011).9. I. D. Karachentsev, Astrophysical Bulletin , 123 (2012).10. R. B. Tully, L. Rizzi, E. J. Shaya, et al., Astronom. J. , 323 (2009).11. I. D. Karachentsev and D. I. Makarov, Astrophysical Bulletin , 299 (2008).12. D. I. Makarov and I. D. Karachentsev, Astrophysical Bulletin , 24 (2009).13. D. I. Makarov and I. D. Karachentsev, Monthly Notices Roy. Astronom. Soc. , 2498 (2011).14. I. D. Karachentsev, D. I. Makarov, V. E. Karachentseva, and O. V. Melnyk, Astrophysical Bulletin ,1 (2011).15. M. Fukugita and P. J. E. Peebles, Astrophys. J. , 643 (2004).16. J. P. Huchra and M. J. Geller, Astrophys. J. , 423 (1982).17. R. B. Tully, E. J. Shaya, I. D. Karachentsev, et al., Astrophys. J. , 184 (2008).18. I. D. Karachentsev, O. G. Nasonova, and H. M. Courtois, Astrophys. J. , 123 (2011).19. R. B. Tully and R. J. Fisher, Astronom. and Astrophys.54