1: %\documentclass[apj]{emulateapj}
2: %\documentclass[apj,onecolumn]{emulateapj}
3: \documentclass[12pt,preprint]{aastex}
4: %\usepackage{amsmath}
5: %\usepackage{txfonts}
6: %\eqsecnum
7:
8: %\documentclass[12pt,preprint]{aastex}
9: %\usepackage{amsmath}
10:
11: %\usepackage{epsf}
12:
13:
14: \renewcommand\email\texttt
15: \newcommand\F[4]{\ensuremath{\mbox{}_2F_1\!\left(#1,#2;#3;#4\right)}}
16: \newcommand\bmath[1]{\ensuremath{\mbox{\boldmath $#1$}}}
17:
18: \newcommand\Teff{{T_{\rm eff}}}
19: \newcommand\FeH{{[{\rm Fe/H}]}}
20: \def\change#1{{#1}}
21:
22: %\newcommand\coords{{14:10:00\, +14:50:00 $\pm 15\arcsec$}} %%%% wrong
23: \newcommand\coords{{14:00:06\, +14:30:00 $\pm 15\arcsec$}} %%%% correct
24: %
25: % \lta and \gta : produce > and < signs with twiddle underneath
26: %
27: \def\spose#1{\hbox to 0pt{#1\hss}}
28: \def\lta{\mathrel{\spose{\lower 3pt\hbox{$\sim$}}
29: \raise 2.0pt\hbox{$<$}}}
30: \def\gta{\mathrel{\spose{\lower 3pt\hbox{$\sim$}}
31: \raise 2.0pt\hbox{$>$}}}
32: \newcommand{\boo}{Bo\"otes\,II}
33: \begin{document}
34:
35: %\slugcomment{\sc submitted to \it Astrophysical Journal Letters}
36:
37: \shorttitle{\sc A Pair of Bo{\"o}tes} \shortauthors{Walsh, Jerjen
38: and Willman}
39:
40: \title{A Pair of Bo{\"o}tes: A New Milky Way Satellite}
41:
42: \author{S.\ M. Walsh\altaffilmark{1},
43: H. Jerjen\altaffilmark{1}, B. Willman\altaffilmark{2}}
44:
45: \altaffiltext{1}{Research School of Astronomy and Astrophysics,
46: Australian National University, Cotter Road, Weston, ACT 2611,
47: \email{swalsh, jerjen@mso.anu.edu.au}} \altaffiltext{2}{Clay Fellow,
48: Harvard-Smithsonian Center for Astrophysics, 60 Garden Street
49: Cambridge, MA 02138}
50:
51: %\newpage
52:
53: \begin{abstract}
54: As part of preparations for a southern sky search for faint Milky
55: Way dwarf galaxy satellites, we report the discovery of a stellar
56: overdensity in the Sloan Digital Sky Survey Data Release 5, lying at
57: an angular distance of only $1.5$\,degrees from the recently
58: discovered Bo{\"o}tes dwarf. The overdensity was detected well above
59: statistical noise by employing a sophisticated data mining algorithm
60: and does not correspond to any catalogued object. Overlaid
61: isochrones using stellar population synthesis models show that the
62: color-magnitude diagram of that region has the signature of an old
63: (12\,Gyr), metal-poor (${\rm Fe/H}\approx-2.0$) stellar population
64: at a tentative distance of $60$\,kpc, evidently the same
65: heliocentric distance as the Bo\"otes dwarf. We estimate the new
66: object to have a total magnitude of $M_{V}\sim-3.1\pm1.1$\,mag and a
67: half-light radius of $r_{h}=4'.1\pm1'.6$ ($72\pm28$\,pc) placing it
68: in an apparent $40<r_{h}<100$\,pc void between globular clusters and dwarf galaxies, occupied only by another recently
69: discovered Milky Way Satellite, Coma Berenices.
70: \end{abstract}
71:
72: \keywords{galaxies: dwarf --- Local Group}
73:
74: %\newpage
75:
76: \section{Introduction}
77: The last three years have seen a torrent of new Milky Way (MW)
78: satellites being discovered in the Northern hemisphere, almost
79: doubling the number known prior to 2005: Bo\"otes (Belokurov et
80: al.~2006a), Canes Venatici (Zucker et al.~2006a), Willman 1 (Willman
81: er al.~2005a), Ursa Major (Willman et al.~2005b), Ursa Major II
82: (Zucker et al.~2006a), Hercules, Coma Berenices, SEGUE 1, Canes
83: Venatici II, Leo IV (Belokurov et al.~2006b), and Leo T (Irwin et
84: al.~2007). Eight of these new objects are consistent in size and
85: luminosity with dwarf spheroidal satellites while Willman 1 and
86: Segue 1 straddle the intersection of dwarfs and globular clusters.
87: Coma Berenices falls in an apparent void of objects spanning the
88: $40-100$\,pc range (see Fig 1. Gilmore et al.~2007).
89: %\cite{bootes}, \cite{cvn}, \cite{willman1}, \cite{uma},
90: %\cite{uma2}, \cite{quintet}, \cite{leot}).
91: These objects were all initially detected as overdensities of
92: resolved stars in the photometric data of Sloan Digital Sky Survey
93: (SDSS, York et al.~2000) and, with the exception of Leo IV,
94: subsequently confirmed with follow-up observations. The numerous
95: discoveries of extremely low surface brightness dwarf spheroidals
96: (dSphs) in the ~1/4 of the sky covered by SDSS strongly suggests
97: that there are many more yet undiscovered. A significant new
98: population of dwarf satellite galaxies would go a long way to
99: reconcile the current discrepancy between Lambda Cold Dark Matter
100: theory predictions (Klypin et al.~1999; Moore et al.~1999) and
101: actual observed dSph numbers.
102:
103: The next few years will see the advent of digital surveys that will
104: enable all-sky searches for Milky Way satellites (e.g. PanSTARRS,
105: Kaiser et al.~2005). We intend to blindly scan the entire Southern
106: sky (20,000 square degrees) for new MW dwarf satellites with the
107: upcoming Southern Sky Survey performed with the 1.3 meter ANU
108: SkyMapper telescope at Siding Spring (Keller et al.~2007). The final
109: combined $\sim25$\,TB catalog of the survey is estimated to reach a
110: signal-to-noise of 5 at $r=22.6$, 1.0 mag deeper than SDSS. In
111: preparation for this survey we are testing sensitive data mining
112: algorithms and search strategies using the freely available SDSS
113: Data Release 5 (DR5, Adelman-McCarthy et al.~2006). A full overview,
114: results and a detailed discussion will be presented in a subsequent
115: paper (Walsh et al. in prep). Our software test on SDSS data has
116: yielded several promising candidates, the most prominent of which we
117: present here. While follow-up observations will reveal the
118: candidate's true nature in more detail, its size and luminosity are
119: consistent with those of the other recent detections that have been
120: labeled dwarfs spheroidals. We thus follow convention and designate
121: it Bo\"{o}tes II.
122:
123: \section{Data and Discovery}
124:
125: DR5 includes a five color photometric catalogue covering 8000 square
126: degrees around the north Galactic pole (Adelman-McCarthy et
127: al.~2006). We have searched this publicly available data for
128: concentrations of old stars at various distance intervals out to the
129: Galactic virial radius of 250\,kpc. We use a method similar to that
130: described in Willman et al. (2002) and Willman (2003), and described
131: in full in Walsh et al. (in preparation). We use a complicated set of cuts to identify all stellar sources in
132: fields of $3^{\circ}$ height in Declination and of arbitrary width
133: in Right Ascension that are consistent in $(g-r,r)$ parameter space
134: with that of a dSph at a desired distance (Red Giant Branch, Blue Horizontal Branch and Main-Sequence Turnoff). We then convolve the
135: binned spatial positions of these sources with an exponential surface
136: brightness profile and subtract the $0.9^{\circ}\times0.9^{\circ}$ running mean from each $0.02^{\circ}\times0.02^{\circ}$ pixel. A density
137: threshold in standard deviations above the local mean is defined as a function of the
138: background stellar density for each pixel, allowing us to search
139: over fields with stellar density gradients. The process is repeated
140: for different magnitude bins to change sensitivity with distance.
141:
142:
143:
144: Applying this method to DR5 we recovered all of the recently
145: reported dSphs, as well as many previously known objects such as
146: globular clusters and background galaxy clusters. The detection
147: significance of the faintest dwarfs, quantified by the parameter
148: $P$ (maximum level above threshold density times area above threshold), are shown in Fig.\,\ref{fig:hist}, along with the result for
149: the new object. The solid line shows the number of ``detections'' in
150: thirty-nine 1000 square degree randomized stellar fields each of
151: varying stellar density to determine foreground contamination from
152: random clustering. The newly discovered satellites (minus SEGUE 1
153: which fell outside the analyzed area) as well as the \boo\,
154: overdensity are all well above the threshold $P=85$ above which our
155: detection algorithm statistically yields less than one false
156: positive detection over the entire area of DR5. The Bo\"otes II
157: overdensity is not associated with any known Galactic or
158: extragalactic object and is consistent in $(g-r,r)$ and
159: size-luminosity space with a new dwarf. Figure \ref{fig:pair} shows
160: the position of Bo\"otes\,II relative to Bo\"{o}tes.
161:
162:
163: % which allows to exclude a background cluster as origin
164:
165:
166: Fig \ref{fig:boo2} shows our detections of Coma Berenices, \boo\,
167: and Bo\"otes. The contours represent the level above the threshold
168: density, which is then multiplied by the detection area to give $P$.
169: Coma Berenices and \boo\,peak at higher densities because they are
170: more concentrated than Bo\"otes, but the latter's spatial extent
171: means that it is still a stronger detection. The detection of \boo\,
172: is consistent in all respects with the detections of the other
173: Galactic satellites, albeit much fainter.
174:
175: \section{Candidate Properties}
176: We use SDSS data to extract as much information as possible and estimate preliminary parameters for \boo. The overdensity is
177: visible even before smoothing and the lack of a concentration of
178: background galaxies (Figure \ref{fig:fits}) allows to exclude a
179: galaxy cluster as an origin. Looking at the CMD in Figure
180: \ref{fig:boo2} reveals a weak red giant branch and blue horizontal
181: branch (or red clump) at a distance modulus apparently identical to
182: that of Bo\"{o}tes, and similar to the Coma Berenices dwarf
183: ($m-M=18.2$, Belokurov et al.~2006b). The CMD features become even
184: more prominent in the associated area-normalized field-subtracted
185: Hess diagram. Overplotted is the isochrone of a metal-poor
186: ([Fe/H]$=-2.0$), old (12\,Gyr) stellar population (Girardi et
187: al.~2004) to illustrate the consistency of our object with an old
188: stellar population. Using the assumed distance modulus of
189: $(m-M)=18.9$ (60\,kpc) which is the same heliocentric distance of
190: the Bo\"otes dwarf, \boo\, would lie at a spatial distance of only
191: $\sim1.6$\,kpc from Bo\"{o}tes. This hints at a physical connection
192: between the two systems, although a further discussion of this idea
193: is beyond the scope of this letter.
194:
195:
196:
197: Figure \ref{fig:sb} presents the azimuthally averaged stellar
198: density profile generated from all stars with $(g-r)<0.65$ and
199: $17.0<r<22.5$ centered on \boo. These data were then fitted with a
200: Plummer profile (dotted line) plus a constant (dashed line), the
201: latter to account for the foreground screen of Galactic stars. The
202: best-fitting profile has a half-light radius of $4.1\pm1.6$\,arcmin,
203: approximately one third of the physical size of Bo\"{o}tes.
204: Alternatively, fitting an exponential profile to the data gives a
205: half-light radius of $4.0\pm1.9$\,arcmin.
206:
207:
208:
209: We use two methods to empirically derive the total magnitude of our
210: object. Firstly we use SDSS coverage of the Draco dSph to calculate
211: the flux ratio of the integrated luminosity functions. We derive a
212: flux ratio $f_{Draco}/f_{Boo\,II}\simeq 172\pm38$. This converts
213: into a magnitude difference of $2.1\pm0.3$\,mag and a total absolute
214: magnitude of $M_{V}\sim -3.8\pm0.6$\,mag for \boo\,, adopting
215: $M_{V}=-9.4$\,mag for Draco (Grebel et al.~(2003)). The same
216: analysis yields $M_{V}\sim -6.0\pm0.6$ for Bo\"{o}tes, 0.7\,mag
217: brighter than $M_{V}\sim -5.3\pm0.6$ given by Belokurov et
218: al.~(2006a). Secondly, we use the integrated surface brightness
219: profiles of Bo\"otes and \boo. The flux ratio from this method gives
220: $f_{Boo}/f_{Boo\,II}\simeq 15$. Using $M_{V}\sim -5.3\pm0.6$ for
221: Bo\"otes gives $M_{V}\sim -2.4\pm0.6$. We therefore adopt a result
222: of $M_{V}\sim -3.1\pm1.1$.
223:
224:
225:
226:
227: \section{Discussion and Conclusion}
228: We report a new Galactic satellite called \boo, only $\sim1.5$
229: degrees away from the Bo\"{o}tes dwarf. This object was discovered
230: as a resolved stellar overdensity in an automated search of SDSS
231: DR5. Any object that is detected by our algorithm will fall in one
232: of the following categories: random foreground clustering, galaxy
233: clusters, stellar associations of partially resolved nearby
234: galaxies, globular clusters, or Galactic dwarf spheroidals. Random
235: clustering at this level is extremely unlikely with $\sim0.008$ such false objects occuring in the entire DR5 area (Fig.~\ref{fig:hist}). The CMD shows an apparent MST and RGB structure that is unlikely to be associated with a distant galaxy cluster and no
236: evidence is found of a suspicious accumulation of background
237: galaxies (Fig.~\ref{fig:fits}).
238:
239: The combined evidences from CMD, surface brightness profile and the
240: good agreement with the size-luminosity relationship of other MW
241: satellites leads us to conclude this object is a previously
242: undiscovered companion to the Milky Way at a tentative distance of
243: $60$\,kpc. But is it a GC or a dSph? Traditionally, these objects
244: are distinguished by their differing physical size; dark matter
245: dominated dSphs are more extended than a purely stellar system of
246: equal luminosity. Equipped with the two parameters $\log(r_h/{\rm
247: pc})=1.8^{+0.2}_{-0.3}$ and $M_{V}=-3.1\pm1.1$\,mag for \boo\, we
248: add our object to the other recently discovered dwarfs in the
249: size-luminosity plot (Fig.~\ref{fig:lr}). The distinction between
250: dSphs and GCs is blurred in the low luminosity regime as was
251: emphasized by the discovery of Willman 1 (Willman et al.~2005a).
252: \boo\, falls alongside Coma Berenices in the 40--100\,pc region devoid of other objects, between globular clusters and dwarfs.
253: \boo\, is of comparable luminosity to SEGUE 1, but is a factor of
254: $\sim2$ larger.
255:
256: With a physical half-light size that is an order of mag larger than
257: most GCs but similar to those of dSphs, we are inclined to designate
258: \boo\, a dwarf galaxy. However, without kinematic data, it is not
259: possible to say for certain whether or not \boo\, formed inside of a
260: dark matter halo, which would confirm it as such. It is also not
261: possible with the current data to determine the extent to which
262: tidal effects may have shaped the observed size and luminosity of
263: \boo. A combination of follow-up spectroscopy and deep imaging will
264: not only enable a more robust evaluation of the dark matter content
265: of this object, but will also enable an evaluation of a possible
266: physical relationship between Bo\"otes and \boo.
267:
268: \acknowledgments We thank Pavel Kroupa, Manuel Metz and Jose Robles
269: for useful discussions and comments in preparation of the
270: manuscript.
271:
272: Funding for the SDSS and SDSS-II has been provided by the Alfred P.
273: Sloan Foundation, the Participating Institutions, the National Science
274: Foundation, the U.S. Department of Energy, the National Aeronautics and
275: Space Administration, the Japanese Monbukagakusho, the Max Planck
276: Society, and the Higher Education Funding Council for England. The SDSS
277:
278: %\bibliographystyle{apj}
279: %\bibliography{paper}
280: \begin{thebibliography}{}
281:
282: \bibitem[Adelman-McCarthy et al.(2006)]{Am06}
283: Adelman-McCarthy, J.~K., et al.\ 2006, \apjs, 162, 38
284:
285: \bibitem[Belokurov et al.(2006b)]{Be06b} Belokurov, V. et
286: al. 2006b, \apjl, 647, L111 %Bootes
287:
288: \bibitem[Belokurov et al.(2006a)]{Be06a} Belokurov, V. et
289: al.\ 2006a, \apjl, 642, L137
290:
291: \bibitem[Gilmore et al.(2007)]{Gi07} Gilmore, G. et al.\ 2007,
292: astro-ph/0703308
293:
294: \bibitem[Grebel et al.(2003)]{Ge03} Grebel, E., Gallagher, J., \&
295: Harbeck, D.\ 2003, \aj, 125, 1926
296:
297: \bibitem[Irwin et al.(2007)]{Ir07} Irwin, M. et al.\ 2007, \apjl,
298: 656, L13
299:
300: \bibitem[Kaiser et al.(2005)]{Ka05} Kaiser, N. et al.\ 2005, Bulletin of the American Astronomical
301: Society, Dec. 2005
302:
303: \bibitem[Keller et al.(2007)]{Ke07} Keller, S. et al.\ 2007, astro-ph0702511
304:
305: \bibitem[Klypin et al.(1999)]{Kl99} Klypin, A., Kravtsov,
306: A.~V., Valenzuela, O., \& Prada, F.\ 1999, \apj, 522, 82
307:
308: %\bibitem[Mateo (1998)]{Ma98} Mateo, M.\ 1998, \araa, 36,435
309:
310: \bibitem[Moore et al.(1999)]{Mo99} Moore, B., Governato, F., Lake, G.,
311: Quinn, T., Stadel, J., \& Tozzi, P.\ 1999, \apjl, 524, L19
312:
313: \bibitem[Walsh et al.(2007)]{Wa07} Walsh, S., Jerjen, H., Willman,
314: B.\ 2007, in preparation
315:
316: \bibitem[Willman (2003)]{Wi03} Willman, B.\ 2003, PhD Thesis,
317: University of Washington
318:
319: \bibitem[Willman et al. (2005a)]{Wi05a} Willman, B., et al. 2005a,
320: AJ, 129, 2692 %Willman 1
321:
322: \bibitem[Willman et al. (2005b)]{Wi05b} Willman, B., et al. 2005b,
323: ApJ, 626, L85 %UMa
324:
325: \bibitem[York et al. (2000)]{Yo00} York D.G., et al. 2000,
326: AJ, 120, 1579 %SDSS
327:
328: \bibitem[Zucker et al.(2006a)]{Zu06a} Zucker, D.~B., et al.\ 2006a,
329: \apjl, 643, L103 %CVn
330:
331: \bibitem[Zucker et al.(2006b)]{Zu06b} Zucker, D.~B., et al.\ 2006b,
332: \apjl, 650, L41 % UMaII
333: \end{thebibliography}
334:
335: \clearpage
336:
337: \begin{deluxetable}{lrlr}
338: \tablewidth{0pt}
339: \tablecaption{\label{tbl:pars}}
340: \tablehead{
341: \colhead{Parameter} &
342: \colhead{Bo\"{o}tes II}
343: }
344: \startdata RA (h m s) &13 58 00 & {[Fe/H]} & $-2.0$ \\
345: Dec (d m s)&+12 51 00 & Age (Gyr) & 12 \\
346: $(l,b)$ & (353.7,68.9) & $\mu_{0,\rm V}$ (Plummer)& $29.8\pm0.8$ \\
347: $(m-M)$& $18.9\pm0.5$ & $r_{h}$ (Plummer) & $4'.1\pm1'.6$\\
348: Distance (kpc) & $60\pm10$ & $\mu_{0,\rm V}$ (Exponential)& $29.6\pm0.8$ \\
349: $M_{V}$ (mag) & $-3.1\pm1.1$&$r_{h}$
350: (Exponential)&$4'.0\pm1'.9$
351: \enddata
352: \end{deluxetable}
353:
354: \clearpage
355:
356: \begin{figure}
357: \plotone{f1.ps}
358: \caption{Curve of false positive detections (black
359: line) as a function of the detection parameter $P$ provided by our
360: data mining algorithm. The detection threshold (vertical dashed
361: line) is located at $P=85$ above which our detection algorithm
362: statistically yields less than one false detection over the entire
363: 8800 square degrees of SDSS-DR5. Overplotted are the detection
364: levels for the faintest known dSphs as well as the Bo\"otes II
365: object. With $P=119$, the Bo\"otes II overdensity is detected just
366: below the UMa II ($P=128$) and UMa ($P=137$) dwarfs.
367: \label{fig:hist}}
368: \end{figure}
369:
370: \clearpage
371:
372: \begin{figure}
373: \epsscale{.70}
374: %\plotone{f2.ps}
375: \plotone{f2small.ps}
376: \caption{Positions of all SDSS stars with
377: $17<r<23$\,mag and $g-r<0.65$. The Bootes dwarf is clearly visible
378: (210.05d +14.50) as is our candidate $1.5$\,degrees to the
379: south-west at 209.55d +12.85.} \label{fig:pair}
380: \end{figure}
381:
382: \clearpage
383:
384: \begin{figure}
385: \epsscale{1.0}
386: \plotone{f3a.ps}
387: \plotone{f3b.ps}
388: \plotone{f3c.ps}
389: \caption{From top to bottom: Coma Berenices, \boo, Bo\"otes. {\em Left Panels}:
390: positions of SDSS stars passing the photometric selection criteria.
391: {\em Middle Left}: Smoothed positions with contours at 0.5,0.75,1.0
392: (thick line),1.2,1.4,1.6 and 1.8 multiples of threshold density.
393: {\em Middle Right}: CMD of region within the 1.0 contour. {\em Right
394: Panels}: field subtracted Hess diagrams of same regions with
395: overlayed stellar isochrone.\label{fig:boo2}}
396: \end{figure}
397:
398: \clearpage
399:
400: \begin{figure}
401: \plotone{f4.eps}
402: \caption{SDSS image ($15\prime\times15\prime$) centered on the detection of
403: Bootes II with SDSS galaxies overlayed (triangles). Left is East and Up is North.}
404: \label{fig:fits}
405: \end{figure}
406:
407: \clearpage
408:
409: \begin{figure}
410: \plotone{f5.ps}
411: \caption{Radial stellar density profile generated from all stars
412: with $(g-r)<0.65$ and $17.0<r<22.5$ centered on the new object (diamonds).
413: The dotted line is a Plummer profile with a scale parameter $a=5.3$\,arcsec and
414: the dashed line is the contribution of foreground stars. The solid line is the combined fit.}
415: \label{fig:sb}
416: \end{figure}
417:
418: \clearpage
419:
420: \begin{figure}
421: \plotone{f6.ps}
422: \caption{Absolute magnitude ($M_{V}$) versus half-light radius
423: ($r_h$ in pc) for all Galactic companions discovered in SDSS.
424: Bo\"otes II is shown as an asterisk, with associated error bars.
425: Lines of constant central surface brightness are shown with
426: $\mu_{V}=27, 28, 29, 30, 31$ and $32$\,mag\,arcsec$^{-2}$ from left to right.}
427: \label{fig:lr}
428: \end{figure}
429:
430: \end{document}
431: