Discovery of a New, Polar-Orbiting Debris Stream in the Milky Way Stellar Halo
aa r X i v : . [ a s t r o - ph . GA ] J un Discovery of a New, Polar-Orbiting Debris Stream in the MilkyWay Stellar Halo
Heidi Jo Newberg1, Brian Yanny2, & Benjamin A. Willett1
ABSTRACT
We show that there is a low metallicity tidal stream that runs along l =143 ◦ in the South Galactic Cap, about 34 kpc from the Sun, discovered fromSEGUE stellar velocities. Since the most concentrated detections are in the Cetusconstellation, and the orbital path is nearly polar, we name it the Cetus PolarStream (CPS). Although it is spatially coincident with the Sgr dwarf trailingtidal tail at b = − ◦ , the metallicities ([Fe/H] = -2.1), ratio of blue stragglerto blue horizontal branch stars, and velocities of the CPS stars differ from Sgr.Some CPS stars may contaminate previous samples of Sgr dwarf tidal debris.The unusual globular cluster NGC 5824 is located along an orbit fit to the CPS,with the correct radial velocity. Subject headings:
Galaxy: structure — Galaxy: halo — Stars: kinematics
1. Introduction
Over the past decade, spatial substructure from tidally disrupted satellites has beendiscovered in the Milky Way’s stellar spheroid (Newberg et al. 2002; Belokurov et al. 2006;Juri´c et al. 2008; Grillmair 2008). However, we expect that a more detailed accretion historyof the Milky Way can be assembled by including the kinematics of the stars (Harding et al.2001), since the kinematic signature of each tidally disrupted satellite remains long after thespatial density has become such a small fraction of the stellar halo density that the streamcannot be identified from spatial information alone.Recently, Yanny et al. (2009b) noticed a co-moving population of low metallicity bluehorizontal branch stars (BHBs) with positions and velocities near, but not coincident with,the Sagittarius dwarf spheroidal trailing tidal stream in the South Galactic Cap, while study-ing spectra of Milky Way stars from the Sloan Digital Sky Survey (SDSS; York et al. 2000) Dept. of Physics, Applied Physics and Astronomy, Rensselaer Polytechnic Institute Troy, NY 12180 Fermi National Accelerator Laboratory, P.O. Box 500, Batavia, IL 60510 l, b ) = (140 ◦ , − ◦ ) and at a distance of 34 kpc from the Sun, with a line-of-sight, GalacticStandard of Rest velocity v gsr = −
50 km s − and metallicity [Fe/H] ∼ -2.0 (see figures 13and 17 of Yanny et al. 2009b). In this paper, we explore the extent and kinematics of thisnew stream.
2. Observations and Data Analysis
We first identify SDSS/SEGUE data release 7 (DR7; Abazajian et al. 2009) spectrathat are likely to be associated with the new stream. Figure 1 shows a color-magnitudediagram of all stellar objects in the South Galactic Cap with essentially zero proper motionand surface gravities of giant stars. Most of these stars are members of the stellar halo.Circled observations have the velocity and metallicity we expect for the new tidal stream.SDSS/SEGUE data has very complex criteria for selecting the stars for spectroscopicobservation, so structure in the distribution of stars in Figure 1 is dominated by selectioneffects. Since it is not possible to know the velocity of a star and it is difficult to determinethe metallicity of a star before the spectrum is obtained, the selection is blind to these twoquantities; therefore substructure can be identified by looking for regions of Figure 1 inwhich the ratio of circled to uncircled points is high. The three boxes labeled blue horizontalbranch (BHB; − . < ( g − r ) < . , . < ( u − g ) < . , . < g < . − . g − r ) + 25 . < g < − . g − r ) + 27 .
12, 16 . < g < . . < ( g − r ) < . , . < g < .
7) in Figure 1 havea relatively high fraction of stars likely to be in the tidal stream, and comparison with M92and M3 fiducials (An et al. 2008), shifted to 34 kpc, shows that they are also likely to befrom the same stellar population. From the BHB fiducials we extracted from the An et al.(2008) data and distance moduli from Harris (1996), we estimate the absolute magnitudeof the BHBs in the color range − . < ( g − r ) < − .
2, where most of the BHBs lie, is M g = 0 . ◦ < Λ ⊙ < ◦ ). The dashedoutline shows velocities of stars in the new stream. At higher Galactic latitude we relied 3 –primarily on the locus of low metallicity RGBs to select the velocities of stars in the newstream. The new stream has lower metallicity than those of the Sagittarius trailing tidaltail, as demonstrated by the fraction of larger to smaller, point-like symbols within the upperoutlined region compared with the lower region with Sgr velocities.We explore the distance to the tidal stream in the lower panel of Figure 2, which shows g vs. b for the stars in the upper plot that are likely stream members, and photometricallyselected BHBs in the region of the newly identified Cetus Polar Stream (CPS). We find anapproximately linear relationship between g and Galactic latitude ( g = − . b + 17 . M g = 0 .
45. Distanceestimates are tabulated in Table 1, with only statistical errors included. Distance errorsmay be systematically too high or too low by 10%, depending on the determination of theabsolute magnitude of BHBs (Sirko et al. 2004).The four panels of Figure 3 show (upper left) an estimate of the positions of the F turnoffstars in the CPS, and the positions of the photometrically selected BHB stars; (upper right)the ( l, b ) distribution of spectra with colors and magnitudes similar to those in the CPS;(lower left) the distribution of F turnoff stars in Sgr and the CPS, with the stars with CPSvelocities superimposed; and (lower right) the same F turnoff stars with the stars with Sgrstream velocities superimposed. Note that there is an overdensity of photometrically selectedBHB stars that lines up with the background-subtracted F turnoff star overdensity, and theCPS velocity-selected BHB, RGB, and LRGB stars, running approximately along Galacticlatitude l ∼ ◦ . Stars that are velocity selected to be candidate Sgr stream stars follow adifferent path in the sky, along the Sgr dwarf tidal tail as tabulated in Newberg et al. (2003).Although the two streams cross near b = − ◦ , they are about 30 ◦ apart at b = − ◦ . Thelack of significant numbers of colored points at l < ◦ or at l > ◦ (where the Sagittariusstream is located) gives us confidence that this stream is not an artifact, and is distinct fromthe previously identified Sagittarius trailing tail. The Galactic longitude of the stream centerin each of the four SDSS stripes 76, 79, 82, and 86, was estimated by comparing the positionsof the F turnoff stars, photometrically selected BHB stars, and velocity-selected BHB, RGB,and LRGB stars. One sigma errors were also estimated by eye. The stream centers andestimated errors are given in Table 1.To estimate the metallicity of the CPS we histogramed spectroscopically selected candi-date stream stars from Figure 3 that have 105 ◦ < l < ◦ . The peak of the distribution forBHB, RGB, and LRGB stars is [Fe/H] ∼ − .
1. The formal mean metallicity of the stars is[Fe/H] ∼ − . ± .
04. There is a population of LRGB stars that is higher metallicity ([Fe/H] ∼ − . − . ± . σ v ), and the number of spectra at each location. The velocity dispersion for each stripe76, 79, 82, and 86 is computed from the spectra with 120 ◦ < l < ◦ that are shown inthe lower left panel of Figure 3, and tabulated in Table 1. Since the intrinsic SDSS/SEGUEradial velocity errors are about 4 km s − , the intrinsic velocity dispersion of the CPS is about4.5 km s − in stripes 76, 82, and 86, and about 10 km s − in stripe 79.
3. Discussion
The Cetus Polar Stream solves a puzzle long pondered by the authors. In Yanny et al.(2000) we discovered the Sgr dwarf tidal tails along the Celestial Equator, including thetrailing tidal tail in stripe 82. We have always wondered why, in Fig. 3 of that paper, theBHB stars at g = 18 in the South Galactic Cap appear offset in position in the sky fromthe Sgr blue straggler (BS) stars that are two magnitudes fainter. The counts of BHB andBS stars, as defined by Yanny et al. (2000), along southern SDSS stripes 79, 82, and 86 arepresented in Figure 4. From this figure, we determine that many of the SDSS BHB starsin the southern stripes that had previously been attributed to the Sgr trailing tidal tail areactually in the CPS. The ratio of BS (higher surface gravity A-colored stars) to BHB starsvaries amongst globular clusters (i.e. see Figures 12-16 of An et al. 2008), and can be usedas an identifying marker in the halo to help separate two populations with distinct origins orevolutionary histories. From Figure 4, it is clear that the Sgr trailing tidal tail has a muchlarger BS/BHB ratio than the CPS.We fit an orbit to the four CPS locations in Table 1, following the procedure used byWillett et al. (2009) and a spherical halo potential ( q = 1 . ◦ , or ∼
10 kpc), which argues in favor of a dwarf galaxy progenitor,though the low velocity dispersion ( ∼ − ) argues for a diminutive dwarf galaxy or 5 –possibly a globular cluster. No known dwarf galaxies lie close to the best fit orbit, butthe globular cluster NGC 5824, at ( l, b ) = (332 . ◦ , ◦ ) is located within 3 ◦ of the orbit,at a very plausibly correct distance, and has a radial velocity that matches the predictedorbit radial velocity within one sigma. NGC 5824 has a well-populated BHB and a measured[Fe/H]=-1.85. NGC 5824 measurements are taken from Harris (1996). Additionally, the tidaldistortion of NGC 5824 measured by Grillmair et al. (2005) and Leon, Meylan & Combes(2000) is aligned with the CPS orbit. Grillmair et al. (2005) show that NGC 5824 hasa central cusp; this massive globular cluster could have once been a dwarf galaxy core(Georgiev et al. 2009). Alternatively, it could be associated with the dwarf galaxy progenitoror be the sole progenitor.
4. Conclusions
A previously unknown, low metallicity tidal debris stream is identified at l = 143 ◦ and 34 kpc from the Sun in the South Galactic Cap. Although it is spatially coincidentwith the Sgr dwarf trailing tidal tail at b = − ◦ , the metallicities ([Fe/H] = -2.1), ratioof BS/BHB stars, and velocities of the Cetus Polar Stream stars are significantly different.Some BHB stars that have been attributed to the Sgr trailing tidal tail by previous authorsare instead part of the CPS. Both the width of the tidal stream ( ∼
10 kpc) suggests a dwarfgalaxy progenitor, though the velocity dispersion ( σ ∼ − REFERENCES
Abazajian K., et al. 2009, ApJS, 182, 543Allende Prieto, C. et al. 2008, AJ, 136, 2070. 6 –An, D., Johnson, J. et al. 2008, ApJS, 179, 326Belokurov, V., et al. 2006, ApJ, 642, L137Georgiev, I. Y., Hilker, M., Puzia, T. H., Goudfrooij, P., & Baumgardt, H. 2009, MNRAS,647Grillmair, C., Freeman, K., Irwin, M., & Quinn, P. J. 1995 AJ 109, 2553Grillmair, C. J. 2008, arXiv:0811.3965Harding, P., Morrison, H. L., Olszewski, E. W., Arabadjis, J., Mateo, M., Dohm-Palmer,R. C., Freeman, K. C., & Norris, J. E. 2001, AJ, 122, 1397Harris, W. E. 1996, AJ, 112, 1487Ibata, R., Lewis, G. F., Irwin, M., Totten, E., and Quinn, T. 2001, ApJ, 551, 294Ivezi´c, Z., et al. 2000, AJ, 120, 963Juri´c, M., et al. 2008, ApJ, 673, 864Law, D. R., Johnston, K. V., & Majewski, S. R. 2005, ApJ, 619, 807Lee, Y. S. et al. 2008, AJ, 136, 2022.Lee, Y. S. et al. 2008, AJ, 136, 2050.Leon, S., Meylan, G. & Combes, F 2000 AA 359, 907.Newberg, H., Yanny, B., et al. 2002, ApJ, 569, 245Newberg, H., Yanny, B., et al. 2003, ApJ, 596, L191Newberg, H., Yanny, B., Cole, N., Beers, T. , Re Fiorentin, P., Schneider, D., and Wilhelm,R. 2007 ApJ 668, 221.Schlegel, D.J., Finkbeiner, D.P., & Davis, M. 1998, ApJ, 500, 525Sirko, E. et al. 2004 AJ 127, 899Wilhelm, R., Beers, T. C., & Gray, R. O. 1999, AJ, 117, 2308Willett, B. A., Newberg, H. J., Zhang, H., Yanny, B., & Beers, T. C. 2009, ApJ 697, 207Yanny, B., Newberg, H. J., et al. 2000, ApJ, 540, 825 7 –Yanny, B., Rockosi, C., Newberg, H. J., Knapp, G. R., et al. 2009a, AJ, 137, 4377Yanny, B., Newberg, H. et al. 2009b, ApJ, in pressYork, D.G. et al. 2000, AJ, 120, 1579
This preprint was prepared with the AAS L A TEX macros v5.2.
Table 1. Cetus Polar Stream Detections l δl b v gsr σ v N d δd ◦ ◦ ◦ km s − km s − kpc kpc144 2 -71 -29.8 6.4 4 36.1 1.9144 3 -62 -42.7 6.5 8 33.8 1.8142 3 -54 -39.2 11.1 16 31.8 1.7142 4 -46 -59.2 5.8 14 30.1 1.6 We show as small black dots the color and g magnitude for all the SDSS DR7 stellar spectra inthe South Galactic Cap ( b < ◦ ) with surface gravities of giant stars (1 < log g < . | µ l | < − , | µ b | < − ). These cuts select objects likely to be in the stellarhalo. The circles show those points that have velocities and metallicities consistent with membership in thenew stellar stream ( − < v gsr < − , − < [Fe / H] < − .
9. Stars with − . < ( g − r ) < . W BG metallicity measurement (blue circles).The stars with 0 . < ( g − r ) < . a metallicity measurement (red circles). Fiducial sequences for M92 and M3, shifted to 34 kpc, are shown forreference. The top panel shows velocities of the stars in the color-magnitude boxes in Figure 1. Spectrawith metallicities of − < [Fe / H] < − . v gsr ∼
150 km s − thatmay have a kinematic association with another (unrelated) structure. We overlay the calculated velocities ofthick disk stars with 50 ◦ < l < ◦ (top to bottom) in magenta to show there is no confusion with streamcandidates even if log g is misidentified. Filled circles in the lower panel show the apparent magnitudes ofthe stars in the upper panel that have metallicities and velocities of the new stream (all circled stars insidethe dashed box in the upper panel, using the same color code). The triangles show photometrically selectedBHB stars (see Yanny et al. 2000 for photometric selection technique) that have 120 ◦ < l < ◦ . The trendof distance with b is consistent between photometrically selected BHBs and those with spectra. Note thatmany of the LRGB stars in the new stream are too faint to be included in the SDSS/SEGUE spectroscopicsurvey, and the intrinsic magnitude distribution of RGB and LRGB stars is very broad. The adopted meanstream velocities (upper panel) and apparent magnitude of the BHB stars (lower panel) are shown as opencircles with errors, and a solid black curve shows the best fit orbit through those stream locations in bothpanels.
10 –
Fig. 3.—
The upper left panel shows the density of turnoff stars in a CPS color magnitude box (20 . 36) minus the density of stars in a Sgr box (20 . < g < . , . < ( g − r ) < . 20) in polar Galactic coordinates, origin at the SGP. The overdensities (dark areas) runningalong l = 140 ◦ from b = − ◦ up to b = − ◦ show the CPS. The blue dots in the upper left panel show thepositions (offset 3 ◦ in b for clarity) of photometrically selected CPS candidate BHBs, with − . b +16 . 24. Note the excess along 130 ◦ < l < ◦ , − ◦ < b < − ◦ . The upper right panelshows (magenta circles) the locations of stars with SDSS/SEGUE spectra in the color-magnitude selectionboxes of Figure 1, showing the completeness coverage of the spectroscopy relative to the imaging. The lowerleft panel shows the density of turnoff stars with 20 . < g < . , . < ( g − r ) < . , ( u − g ) > . 11 –Fig. 4.— We show the number of photometrically selected BHB and BS stars in SDSS stripes 79, 82,and 86, in the South Galactic Cap. The stars are selected from SDSS imaging data with − . < ( g − r ) < . , . < ( u − g ) < . µ ), which is identical to right ascension α for stripe 82, and deviatesfrom αα