1: \documentclass[12pt,preprint]{aastex}
2: %\usepackage{graphicx,calc,caption,fullpage}
3: \usepackage{multirow}
4:
5:
6: \begin{document}
7:
8: \slugcomment{\it Accepted for publication in ApJS on 1 September 2007.}
9:
10: \title{CGRaBS: An All-Sky Survey of Gamma-Ray Blazar Candidates}
11:
12: \author{Stephen E.\ Healey\altaffilmark{1}, Roger W.\ Romani\altaffilmark{1},
13: Garret Cotter\altaffilmark{2}, Peter F.\ Michelson\altaffilmark{1}, \\
14: Edward F.\ Schlafly\altaffilmark{1}, Anthony C.\ S.\ Readhead\altaffilmark{3},
15: Paolo Giommi\altaffilmark{4}, \\
16: Sylvain Chaty\altaffilmark{5}, Isabelle A. Grenier\altaffilmark{5},
17: Lawrence C. Weintraub\altaffilmark{3}}
18:
19: \altaffiltext{1}{Department of Physics/KIPAC, Stanford University, Stanford, CA 94305; sehealey@astro.stanford.edu, rwr@astro.stanford.edu.}
20: \altaffiltext{2}{Department of Astrophysics, University of Oxford, Oxford OX1 3RH, UK}
21: \altaffiltext{3}{Department of Astronomy, California Institute of Technology, Pasadena, CA 91125}
22: \altaffiltext{4}{ASI Science Data Center, I-00044 Frascati, Italy}
23: \altaffiltext{5}{Laboratoire AIM, CEA/DSM - CNRS - Universit\'{e} Paris Diderot,
24: Service d'Astrophysique, CEA Saclay, 91191 Gif/Yvette, France}
25:
26: \begin{abstract}
27: We describe a uniform all-sky survey of bright blazars, selected primarily by their
28: flat radio spectra, that is designed to provide a large catalog of likely
29: $\gamma$-ray AGN. The defined sample has 1625 targets with radio and X-ray properties similar to
30: those of the EGRET blazars, spread uniformly across the $|b|>10^\circ$ sky. We also report
31: progress toward optical characterization of the sample; of objects with known $R<23$, 85\% have been
32: classified and 81\% have measured redshifts. One goal of this
33: program is to focus attention on the most interesting (e.g., high redshift, high luminosity, ...)\
34: sources for intensive multiwavelength study during the observations by the Large Area Telescope
35: (LAT) on {\it GLAST}.
36: %EDIT - CUT
37: %(LAT) on {\it GLAST}. As one of the largest uniform blazar samples, the data
38: %will also be helpful in the study of the evolution of black holes and high-power jets.
39: \end{abstract}
40:
41: \keywords{BL Lacertae objects: general --- galaxies: active --- quasars: general --- surveys}
42:
43: \section{Introduction}
44:
45: It is well-known \citep{3eg,mattox} that many of the high-latitude EGRET sources
46: are associated with the bright, flat radio spectrum AGN known as blazars.
47: Sowards-Emmerd et al.\ (2003; SRM03) quantified such associations, developing a
48: combined figure of merit (FoM), which measured the likelihood that an individual
49: radio/X-ray source near the large ($\sim$$0.7^\circ$) Third EGRET Catalog
50: (3EG, \citealt{3eg}) uncertainty regions is the $\gamma$-ray counterpart.
51: They also noted that there are many radio-loud blazars with very
52: similar properties not obviously associated with a 3EG source. A likely explanation
53: is that blazars are very variable at high energy, with duty cycles for the bright, flaring
54: state as small as a few percent \citep{hart93,kniffen}. During the limited
55: (typically two weeks per pointing direction) 3EG exposure many of these sources may have
56: been in quiescence. Accordingly, \citet{srm05} extended the SRM03 analysis
57: by selecting ``3EG-like'' blazars, i.e., sources whose radio flux density and spectrum
58: (and X-ray flux) were very like those of the 3EG blazars but which happened not to
59: lie within a 3EG test statistic (TS) uncertainty region. The positions of these
60: sources showed a clear excess of $\gamma$-ray photons over background and these sources
61: are likely to show $\gamma$-ray high states during future missions.
62:
63: The Large Area Telescope (LAT) on {\it GLAST} will provide an improvement of several
64: orders of magnitude over EGRET/{\it CGRO} with an increased sensitivity in the 50 MeV $-$ 300 GeV
65: range and a wide ($>$2.5 sr) field of view. The LAT should detect
66: many thousands of sources during the 5- to 10-year mission, with a large fraction of the
67: high-latitude sources being blazars. The early mission will be devoted to a sky survey, covering
68: the entire sky at good sensitivity every three hours. This will greatly
69: enhance the likelihood of detecting transient and variable sources, such as
70: blazars. While several large samples of blazars have been compiled
71: %EDIT
72: %recently (see especially the ASDC blazar catalog and the ROXA catalog, \citealt{asdc}), there is a
73: recently (see especially the ASDC blazar catalog, \citealt{asdc}, and the ROXA catalog,
74: \citealt{roxa}), there is a
75: surprisingly incomplete knowledge of the radio-bright, flat-spectrum population,
76: which is most clearly associated with the GeV $\gamma$-ray sources. We seek
77: to rectify this by defining CGRaBS, the \underline{C}andidate \underline{G}amma-\underline{Ra}y
78: \underline{B}lazar \underline{S}urvey, a large sample of EGRET-like blazars selected across
79: the extragalactic sky. By obtaining optical classifications and redshifts for a large
80: fraction of these sources, we plan to enable prompt, intensive follow-up of the most
81: interesting (e.g., high redshift, high luminosity, peculiar spectrum) sources that {\it are}
82: detected in the LAT sky survey data. Furthermore, identification of a
83: substantial fraction of the LAT sources with blazars will allow us
84: to focus on the non-blazar remainder, potentially isolating new classes of
85: cosmic $\gamma$-ray emitters.
86:
87: \section{Sample Selection}
88:
89: For any FoM-type counterpart selection, it is important to have uniform
90: parent populations. \citet{crates} have recently developed such a catalog, CRATES,
91: which extended results of the CLASS survey \citep{class} to obtain
92: 8.4 GHz observations of all $|b|>10^\circ$ objects brighter than 65 mJy at 4.8 GHz
93: with spectral indices $\alpha > -0.5$ (where $S \propto \nu^\alpha$). To estimate
94: the radio spectral index of the core, we use the NVSS \citep{nvss} and
95: SUMSS \citep{sumss} lower-frequency surveys. The result is
96: a sample of over 11,000 flat-spectrum radio sources with interferometric measurements
97: at $\sim$1 GHz and 8.4 GHz (with FWHM beam sizes $\sim$$40\arcsec$ and $\sim$$0.25\arcsec$
98: respectively), giving precise positions, spectral indices, and morphologies
99: for the compact components. The CRATES catalog is as uniform as possible for the high-latitude
100: ($|b|>10^\circ$) sky, limited by gaps in which the initial 4.8 GHz data are unavailable.
101: We believe that this catalog is an excellent starting point for
102: comparison with other all-sky samples (e.g., microwave and $\gamma$-ray).
103:
104: Here we wish to find EGRET-like blazars, so we adopt the
105: FoM of SRM03, which was derived from comparing the well-established 3EG
106: blazar sources with the northern (CLASS-generated) subset of
107: the CRATES catalog. This FoM is given by the heuristic fitting formula
108: $\mathrm{FoM_{3EG}}=100 \times P_\alpha \times P_S \times P_{\mathrm X} \times P_{\mathrm{TS}}$, where the $P$
109: terms are ``excess probabilities'' for the observed parameters for radio sources near
110: 3EG sources. Here, $P_\alpha = 0.19-0.35\alpha_{\mathrm{low}/8.4}$ ($0 \le P_\alpha \le 0.4$),
111: $P_S=-3.47+2.45 \log_{10} S_{8.4} - 0.34 \log_{10}^2S_{8.4}$ ($0 \le P_S \le 1$)
112: and $P_\mathrm{X} = 0.99 + 0.41 \log_{10} F$ ($0.5 \le P_\mathrm{X} \le 1$), with $F$ the RASS \citep{rass}
113: counts per second and the $P$ terms bounded to the ranges in parentheses. Finally,
114: $P_{\mathrm{TS}}=1-\mathrm{CL}$, where CL is the confidence limit of the 3EG source
115: localization contour passing through the position of the radio source.
116: In essence, the FoM is
117: composed of the product of the ``excess'' probabilities of sources of a given flux density, spectral
118: index, etc.\ over random chance. While the FoM probability is not directly normalized,
119: ``false positive'' rates were computed at each FoM level by comparison with the statistics of
120: scrambled versions of the sky catalogs.
121: Of course, once we have an initial survey of LAT blazar sources, it will be
122: appropriate to derive new coefficients, ``re-training'' the FoM against this sample.
123:
124: To develop an all-sky survey of blazar candidates, we compute an FoM for
125: each source in the CRATES catalog. We must do this without reference
126: to 3EG sources. Thus for this paper we define $\mathrm{FoM}= P_\alpha \times P_S \times P_\mathrm{X}$.
127: To connect with $\mathrm{FoM_{3EG}}$, note that a blazar with the present $\mathrm{FoM}=0.2$ would
128: correspond to a $\mathrm{FoM_{3EG}}=1$ at the 95\% localization contour of a $\gamma$-ray
129: source, a ``likely'' ($>$90\% correct) identification.
130: With this definition, 5059 of the CRATES sources have a nonzero FoM.
131: To focus our follow-up on the best and most 3EG-like objects, we define CGRaBS as
132: those 1625 sources with $\mathrm{FoM} > 0.04$. This corresponds to an SRM03 value of
133: $\mathrm{FoM_{3EG}}=2$ at the $50\%$ localization contour, a very likely association,
134: and a $\mathrm{FoM_{3EG}}=0.2$ for a source at the 95\% confidence contour, a reasonable
135: ($>$80\%) likelihood of an association.
136: Figure 1 shows an Aitoff equal-area projection of the CGRaBS sample along with its parent
137: survey, CRATES. Figure 2 shows a projection of the CGRaBS sample indicating the FoM of
138: each source.
139:
140: \begin{figure}[h]
141: \begin{center}
142:
143: \includegraphics[width=\textwidth,angle=0,keepaspectratio=true]{f1.eps}
144:
145: \caption{Aitoff equal-area projection of the CRATES parent sample (small dots) and the
146: CGRaBS sample (large dots) in Galactic coordinates $(l,\,b)$. The central meridian is
147: $l = 0^\circ$. A few small holes are visible just below $\delta=0^\circ$ (dot-dashed line),
148: stemming from incomplete PMN sky coverage.}
149: \label{aitoff1}
150: \end{center}
151: \end{figure}
152:
153: \begin{figure}[h]
154: \begin{center}
155:
156: \includegraphics[width=0.8\textwidth,angle=0,keepaspectratio=true,trim=1.4in 0.8in 1in 3.75in]{f2.eps}
157:
158: \caption{Aitoff equal-area projection of the CGRaBS sample in Galactic coordinates $(l,\,b)$.
159: The central meridian is $l = 0^\circ$. The radius of each dot is proportional to the FoM of the
160: source; the dot for a source with FoM = 0.1 is shown for comparison. The dot styles indicate
161: optical classifications (see \S\S3.2.1-2): $\bullet$ = FSRQ, $\otimes$ = BLL, $\oplus$ = AGN,
162: $\circ$ = unknown.}
163: \label{aitoff2}
164: \end{center}
165: \end{figure}
166:
167: The radio spectral index is a major component of our FoM; thus, since the
168: interferometric observations at 8.4 GHz and low frequency were non-simultaneous, variability
169: can, in principle, affect our FoM measurements. Luckily, the variability in the radio
170: is modest compared to the high energy bands: \citeauthor{crates} found that the mean
171: 8.4 GHz variability is $\le$14\% and the low-frequency variability on the relevant
172: several-year timescale is even smaller. Thus, we do not expect that radio variability
173: will dramatically affect our FoM estimates. Further, the (more likely variable) RASS
174: detections turn out not to be a major selection bias in this survey.
175: If the X-ray contribution to the FoM is ignored and a purely radio-based FoM
176: is computed, then 98.5\% of our sources still satisfy the CGRaBS FoM cutoff. Thus, while
177: the X-ray flux from a small number of sources boosts them into the sample, the main effect
178: of the X-ray contribution is to shuffle the ranking within the set of sources that are
179: already qualified. Since the radio FoM weighting increases for bright and inverted (rising)
180: spectra, its net effect is to impose an effective {\it extrapolated} flux density limit at
181: a higher radio frequency. For example, the FoM = 0.04 cutoff corresponds to an
182: extrapolated flux density at 100 GHz of $S_{100} > 230$ mJy (although we do not
183: expect all sources to have a constant $\alpha$ to such high frequency). Less than
184: 1\% of the full CGRaBS targets have an extrapolated flux below this threshold,
185: and these are all low-FoM sources with very high
186: X-ray flux (i.e., largely high-peaked sources; see the next section).
187:
188: Three CGRaBS sources warrant special comment. The CRATES entry for J0352$-$2514 is
189: a combination of 8.4 GHz observations at two epochs, one with an unflagged mapping error and a grossly
190: erroneous position. The CGRaBS entry for J0352$-$2514 uses only the good epoch to determine the correct
191: position, the 8.4 GHz flux density, and the spectral index. Sources J0805$-$0111 and J1639$+$1632 have
192: nominal CGRaBS spectral indices (and thus FoMs) that are almost certainly overestimates. Their NVSS
193: counterparts have marginally resolved jet structure, and the NVSS decompositions offset the core toward
194: the jet. A faint, spurious counter-jet component was introduced and, being slightly closer to the 8.4
195: GHz position, was selected as the 1.4 GHz counterpart, leading to a highly inverted spectral index and a high FoM.
196: We include these sources in the survey since they satisfy the CGRaBS prescription; a more careful treatment of
197: the NVSS counterparts would give a smaller spectral index and FoM. This effect is quite rare, occurring in CGRaBS
198: for only these 2 sources (out of 1625, or 0.12\%) and in CRATES for no more than 20 sources (out of 11,131,
199: or 0.18\%).
200:
201:
202:
203: \section{Optical follow-up}
204:
205: We have specifically {\it not} required a previous optical (or X-ray) detection of
206: our blazar candidates. This radio-driven selection allows us to sample completely the
207: flat-spectrum sources and avoid biasing the detected population.
208: For example, X-ray--bright sources are preferentially low-power ``blue'' blazars
209: such as BL Lacs (so-called high-peaked blazars, or HBLs; \citealt{hbl}). Similarly, requiring
210: optically bright counterparts can bias
211: the sample toward low redshift. However, since the principal goal of the CGRaBS project is
212: to secure optical identifications, we do need good magnitude estimates. To
213: maximize uniformity, we are working toward complete identification for $R < 23$.
214: In practice, we have also observed a number of radio-bright and
215: X-ray/$\gamma$-ray--bright but optically faint sources beyond this limit to explore the
216: extrema of the population.
217:
218: \subsection{Counterparts and photometry}
219:
220: One defining blazar characteristic is rapid optical variability. Thus, we must set a fiducial ``epoch''
221: for the optical magnitudes. In practice, we use the USNO-B1 catalog \citep{usno}
222: since this is the largest source of suitable $R$ magnitudes; we take the fiducial magnitude
223: to be that of the more sensitive second epoch survey (R2). Since we have precise radio
224: positions for the cores of all sources, we identify a USNO-B1 source as the
225: counterpart of a CGRaBS source if the optical position is within $1.5\arcsec$ of the radio
226: position. This gives a large fraction of the required magnitudes, with completeness dropping
227: between $R\sim 20$ and $R\sim21$.
228: For the north Galactic cap, we can supplement these with SDSS identifications
229: (through Data Release 5, \citealt{sdss}) to $r^\prime \la 22$. In
230: confused cases, these archival data were examined visually to determine the best counterpart
231: match. In a number of cases, we were also able to see clear counterparts that were
232: too faint for inclusion in the USNO-B1 catalog but whose magnitudes could be reasonably estimated
233: by measurement of the digitized plate data. In view of the variable blazar
234: magnitudes and non-stellar colors, this low-precision photometry is adequate
235: to guide the follow-up spectroscopy.
236:
237: To complete the process of optical identification (and to improve a few poor USNO-B1
238: magnitudes), we have conducted our own imaging campaign, primarily at the 5 m Hale
239: Telescope at Palomar, the 3.6 m New Technology Telescope (NTT) at La Silla, and the 2.7 m
240: Harlan J.\ Smith Telescope at McDonald. Typical
241: exposures were 180 s through Gunn $r^\prime$ under varying conditions, and magnitudes were
242: calibrated against multiple field stars. For some particularly
243: interesting sources (e.g., high radio-to-optical flux ratio candidates for high redshift),
244: these were supplemented with {\it izJHK} imaging. We do not report here on
245: these optical/IR SEDs. All $r^\prime$ magnitudes have been converted to $R$
246: using the average color term ($R-r^\prime=- 0.253$) of CGRaBS sources detected by
247: both the SDSS and USNO-B1. A magnitude (or limit) for each source is listed in Table 2.
248: For some of the lowest redshift sources, the magnitude is dominated by the flux from
249: the (extended) host galaxy. We also list the nominal Galactic extinction for
250: the source direction $A_R$, derived from the \citet{schleg} maps. Even though the sources are at
251: high latitude, there are a few targets behind dust clouds, indicating a large nominal
252: extinction. However, we do not expect extinction to bias our measured population as
253: the large $A_R$ are not preferentially associated with the faint targets. Furthermore,
254: only 4\% of the blazars have $A_R>1$ and 0.5\% have $A_R>2$; only 4 sources are
255: excluded from the targeted $R=23$ sample by the known extinction.
256: As of 2007.5, there are 88 objects
257: (5.4\%) that do not have measured $R$ magnitudes; of these, 45 have limits fainter than
258: $R=23$ and thus do not nominally require spectroscopy for the complete survey. The sources
259: with brighter limits will be the subject of further imaging. Note that with 68 CGRaBS
260: sources known to be fainter than $R=23$, we expect that the survey will be $>$95\% complete at this
261: magnitude limit.
262:
263: Figure 3 shows the distribution of $R$ magnitudes and limits.
264: Since the $A_R$ are in general small,
265: the extinction-corrected histogram is very similar. At first sight,
266: the rapid drop between $20<R<21$ would seem to be due to the USNO-B1 survey completeness limit.
267: However, we have sufficient deeper CCD imaging to determine that the drop in numbers is largely
268: intrinsic, although we need to complete the imaging before we can characterize the details
269: of the faint source distribution. The right panel shows that we need to complete
270: identifications to faint magnitudes ($R>19$) to get a representative sample of the higher redshift
271: sources.
272:
273: \begin{figure}[ht]
274: \centering
275: \plottwo{f3a.eps}{f3b.eps}
276: \label{magplots}
277: \caption[magplots]{{\it Left:} Magnitude distribution of CGRaBS sources. Lower limits on $R$ are
278: shown by the dashed histogram. {\it Right:} Magnitude distributions for low-redshift and
279: high-redshift sources.}
280: \end{figure}
281:
282:
283: \subsection{Spectroscopy}
284:
285: Our spectroscopic goals are a basic classification of the AGN type, redshift measurement,
286: and measurement of emission line equivalent strengths and kinematic widths (for luminosity
287: and mass evolution studies). Thus, the bulk of our new observations have been low-resolution
288: $\mathcal{R} \sim 500$$-$1500 long-slit spectroscopy. Most of the sources are flat-spectrum
289: radio quasars whose broad lines allow easy identification with relatively low signal-to-noise.
290: However, a significant fraction of the sources ($\sim$15\%) are weak-lined BL Lac sources. For
291: these, we require high S/N and/or high resolution to determine the redshift from host
292: absorption lines. Such measurements require long exposures with large telescopes. At present,
293: %EDIT
294: %we have identified sources as BL Lacs to $R\sim 20$, but our ability to solve for the redshift
295: we have identified sources as BL Lacs to $R\sim 20$, but our ability to measure the redshift
296: drops significantly above $R\sim 18.5$; these BL Lacs are the subject of further spectroscopy
297: at higher dispersion. In this paper, we present a progress report on the optical identifications.
298: Additional papers will discuss the properties of the complete sample, the source SEDs,
299: and the constraints on blazar evolution.
300:
301: \subsubsection{Observations}
302:
303: A fair fraction of the CGRaBS sources are bright, well-known AGN; thus, we
304: have vetted our catalog against the twelfth edition of the V\'{e}ron quasar catalog \citep{veron}
305: and the SDSS DR5 quasar catalog \citep{sdssq}. We have also queried
306: NED ({\tt http://nedwww.ipac.caltech.edu/}) for all CGRaBS sources to find any other redshifts
307: and identifications in the literature.
308: %EDIT - CUT/MOVE
309: %The classification of sources in the literature is a bit heterogeneous. For example, V\'{e}ron
310: %does not distinguish flat-spectrum radio quasars (FSRQs) from other QSOs. All of our
311: %quasar sources, of course, have high radio luminosity and are bona fide FSRQ. Also, this
312: %catalog designates sources
313: %with absolute magnitude $M_V>-23$ as AGNs, not QSOs. To use our $R$ magnitudes
314: %and new redshifts consistently, for sources with $R>15$, we compute $M_R$ for the
315: %$\Lambda$CDM concordance cosmology (smaller $R$ values are usually host-dominated, in any case).
316: %Given the very wide range of observed spectral indices, we did not attempt to apply
317: %any $K$-correction. To summarize: we list here three blazar designations: continuum-dominated
318: %objects are called BLL, strong-lined luminous sources ($M_R<-23$) are labeled FSRQ and
319: %low luminosity sources AGN.
320: %SDSS object type classifications include designations for the criterion for spectroscopic
321: %targeting (e.g. ROSAT, FIRST, etc). We convert such classifications to AGN and FSRQ as appropriate.
322: Archival data identify $\sim$45\% of the CGRaBS objects ($\sim$60\% of the
323: redshifts in hand); the remainder are the targets of our own spectroscopic campaigns.
324: The great workhorse of our spectroscopic effort has been the 9.2 m Hobby*Eberly Telescope (HET)
325: at McDonald, which has observed hundreds of CGRaBS sources in the accessible declination
326: band $-11^\circ < \delta < +73^\circ$. The telescope is fully queue-scheduled \citep{shetrone},
327: allowing us to receive data remotely year-round and to spread the cost of inclement weather and
328: unfavorable conditions. We use the Marcario Low Resolution Spectrograph (LRS; \citealt{lrs})
329: with grism G1 (300 lines mm$^{-1}$), $2\arcsec$ slit, and a Schott GG385 long-pass filter
330: for a resolution of $\mathcal{R} \approx 500$. Typical exposures are $2\times 600$ s,
331: providing redshifts of emission-line objects to $R \sim 22$; brighter objects are
332: also observed under poor conditions with $2\times 300$ s.
333:
334: In addition to our ongoing HET observations, we have mounted dedicated campaigns at
335: a number of other facilities. We conducted three runs totaling 13 nights (over half lost
336: to weather) on the 2.7 m Harlan J. Smith Telescope at McDonald, using
337: the Imaging Grism Instrument (IGI) and the 6000 \AA\ VPH grism. We observed 28
338: objects with the 1.5 m telescope at Cerro Tololo in the ``13/I'' setup (grism 13, 150 lines
339: mm$^{-1}$) in service observing mode as part of the Small and Medium Aperture
340: Research Telescope System (SMARTS) program. We conducted two runs totaling 8 nights on the
341: 3.6 m NTT at La Silla with the ESO Multi-Mode Instrument (EMMI) in the low-resolution
342: spectroscopy (RILD) mode and grism 2 (300 lines mm$^{-1}$). To date, we have had three runs
343: totaling 12 nights on the 5 m Hale Telescope at Palomar with the double
344: spectrograph (DBSP), using a 300 lines mm$^{-1}$ grating on the blue side and a 316 lines
345: mm$^{-1}$ grating on the red side. We observed 12 objects with the 8.2 m Kueyen telescope
346: (the second unit telescope at the Very Large Telescope, VLT) in service observing mode with Focal
347: Reducer/Low-Dispersion Spectrograph 1 (FORS1) and grism GRIS\_300V (300 lines mm$^{-1}$). Finally,
348: we have had three runs totaling 4 nights on the 10 m Keck I Telescope at Mauna Kea (however,
349: the night of 2006 October 28 was the first observing night after the 2006 earthquake, and pointing
350: was severely restricted; observations remained substantially constrained even on the night of
351: 2006 November 24). For these observations, we used the Low Resolution Imaging Spectrometer (LRIS),
352: employing a 600 lines mm$^{-1}$ grism on the blue side and a
353: 300 lines mm$^{-1}$ grating on the red side. A summary of the observations is shown in Table 1.
354:
355: The 1.5 m telescope observations were taken with a fixed N-S slit within a few hours
356: of culmination. For all other systems, observations were taken with a long slit at the
357: parallactic angle. Basic reduction steps were applied to the spectra using standard IRAF
358: routines. Although every effort was made to minimize differential slit losses, in view of the
359: variable slit widths and seeing, we have not attempted to derive absolute spectrophotometry. After
360: standard star calibration, we estimate that the relative spectrophotometric accuracy is $\sim$30\%,
361: based on comparisons of observations of individual targets at different epochs with different
362: instruments. Spectra were corrected for telluric absorption, and all observations for
363: a given target were combined, weighted by S/N, to produce a final spectrum. Sample spectra
364: are shown in Figure 4.
365:
366: \begin{deluxetable}{ccccccc}
367: \tabletypesize{\scriptsize}
368: \tablecaption{Summary of CGRaBS observations.}
369: \tablewidth{0pt}
370: \tablehead{
371:
372: &
373: &
374: \colhead{Wavelength}&
375: \colhead{Spectral}&
376: \colhead{Typical}&
377: \colhead{Typical}\\
378:
379: \colhead{Telescope}&
380: \colhead{Dates}&
381: \colhead{range}&
382: \colhead{resolution}&
383: \colhead{seeing}&
384: \colhead{exposure}\\
385:
386: \colhead{}&
387: \colhead{}&
388: \colhead{(\AA)}&
389: \colhead{(\AA)}&
390: \colhead{(arcsec)}&
391: \colhead{(s)}}
392:
393: \startdata
394: 9.2 m Hobby*Eberly&Ongoing, 2002--present&4100--9700&17&1.5&600, 1200\\[3pt]
395: \hline\\[-6pt]
396: &2005 May 27--31&4250--8250&12&1.5&600, 1200\\
397: 2.7 m Harlan J. Smith&2005 Oct 27--31&4250--8250&12&2.0&600, 1200, 1800\\
398: &2007 Mar 26--28&4250--8250&12&3.0&600, 1200, 1800\\[3pt]
399: \hline\\[-6pt]
400: 1.5 m CTIO&2005B&3500--9000&17&1.5&1200, 1800\\[3pt]
401: \hline\\[-6pt]
402: \multirow{2}{*}{3.6 m NTT}&2006 Aug 29--Sep 1&3900--9100&10&1.3&600, 1800\\
403: &2007 Jan 22--25&3900--9100&10&1.0&600, 1200\\[3pt]
404: \hline\\[-6pt]
405: \multirow{4}{*}{5 m Hale}&2005 Nov 5--9&3300--9500&5\tablenotemark{a}, 16\tablenotemark{b}&2.5&600, 1200\\
406: &2006 Aug 17--18&3300--9500&5\tablenotemark{a}, 16\tablenotemark{b}&1.7&600, 1200\\
407: &2007 Jan 15--16&3300--9500&5\tablenotemark{a}, 16\tablenotemark{b}&2.0&600, 1200\\
408: &2007 Apr 19--21&3300--9500&5\tablenotemark{a}, 16\tablenotemark{b}&2.5&600, 1200\\[3pt]
409: \hline\\[-6pt]
410: 8.2 m VLT-Kueyen&Period 78&3500--8000&17&1.2&600, 900, 1200, 1800\\[3pt]
411: \hline\\[-6pt]
412: &2006 Jul 22--23&3300--9300&3\tablenotemark{a}, 11\tablenotemark{b}&1.5&600, 1200\\
413: 10 m Keck I&2006 Oct 28&3300-9300&3\tablenotemark{a}, 11\tablenotemark{b}&1.5&600, 1200\\
414: &2006 Nov 24&3600-9600&3\tablenotemark{a}, 11\tablenotemark{b}&2.5&600, 1200\\
415: \enddata
416:
417: \tablenotetext{a}{Blue side.}
418: \tablenotetext{b}{Red side.}
419:
420: \end{deluxetable}
421:
422: \begin{figure}[h]
423: \centering
424: \epsscale{0.795}
425: \plotone{f4.eps}
426: \label{sampspec}
427: \caption[sampspec]{Sample CGRaBS spectra.}
428: \end{figure}
429:
430: \subsubsection{Results}
431:
432: Our spectral analysis starts with a basic source classification. The vast majority
433: ($84$\%) are flat-spectrum radio quasars (FSRQs) dominated by strong broad emission lines. The weak-lined
434: BL Lac (BLL) class is somewhat heuristically defined; here we designate as BLLs sources
435: that exhibit the following properties \citep{marcha}: (1) emission line equivalent width $<5$
436: \AA, and (2) H/K 4000 \AA\ break contrast $\equiv (f^+ - f^-)/f^+ < 0.4$, where $f^+$ ($f^-$) is the
437: flux density redward (blueward) of the break. It is often possible to establish that a source is a BLL
438: even when the redshift is impossible to determine. For sources with $R>15$, we compute $M_R$ for the
439: $\Lambda$CDM concordance cosmology (smaller $R$ values are usually host-dominated, in any case) and
440: classify broad emission line sources with $M_R>-23$ as AGN. Thus, we list here three blazar
441: designations: continuum-dominated BLLs, low-luminosity broad-line AGN, and luminous broad-line FSRQs.
442: A small number of non-blazar sources is also present. Sources with narrow lines ($v_\mathrm{FWHM} <
443: 1000$ km s$^{-1}$) are denoted as narrow-line radio galaxies (NLRGs). Sources with small line
444: equivalent widths but large H/K break contrasts are denoted as galaxies. These low-redshift sources may
445: represent the low-luminosity extension of the blazar phenomenon. One extremely compact planetary nebula
446: (PN) made our survey cuts. Finally, in four cases, the radio position was within
447: $1.5\arcsec$ of a field star that dominated the initial spectrum. With improved imaging,
448: the fainter CGRaBS blazar counterparts can be identified.
449:
450: Redshifts were measured by cross-correlation analysis. For a modest number of FSRQs,
451: only a single strong, broad emission line is identified. In most cases, we conservatively identify this as
452: Mg {\sc ii} $\lambda 2800$, supported by the absence of strong lines expected for other
453: identifications and, often, by Fe {\sc ii} structure in the surrounding continuum. Nevertheless, these
454: redshifts are flagged by a colon (:), indicating possible systematic uncertainty. Absorption line
455: redshifts were obtained for some BLLs. In a few
456: cases, the BLL sources had multiple observations, and we were able to obtain emission line
457: redshifts when the source was in a {\it low} continuum state and the emission line
458: equivalent widths were relatively large. A few additional BLLs have redshift constraints,
459: with upper limits from the lack of Ly-$\alpha$ absorption in the UV and lower limits from
460: clearly identified (typically Mg {\sc ii}) intergalactic absorption systems. We have also measured
461: continuum flux densities and equivalent and kinematic widths for the strong optical/UV resonance
462: lines. These will be used to study the black hole masses and evolution.
463:
464: Table 2 presents the first page of the CGRaBS catalog; the full table appears in the online
465: edition. Here we include the precise position, the 8.4 GHz core flux density, the FoM,
466: the $R$ magnitude, the extinction $A_R$, and the optical classification and redshift, if any.
467:
468: \input{tab2.tex}
469:
470: \section{Discussion}
471:
472: To date, we have 1226 redshifts and 64 BLLs with unknown redshift. Thus classification is
473: 79\% complete with respect to the entire survey and 85\% for objects with known $R < 23$.
474: So far, 10.3\% of all
475: %EDIT - added AGN fraction
476: objects classified are BLLs, 3.4\% are AGN, and 1\% are NLRGs. Figure 5 shows the completeness as a
477: function of magnitude. Source classification and redshifts are $>$85\% complete to $R=20$.
478: While the completeness
479: drops off rapidly beyond this, so do the source counts, and so reaching $>$95\% completeness
480: at the survey limit is feasible. Note, however, that only $\sim$52\% of the
481: BLLs have redshifts and that this fraction falls off quickly above $R=18$. Clearly, pushing the
482: largely complete BLL sample fainter than $R=20$ will be a challenge.
483:
484: \begin{figure}[ht]
485: \centering
486: %\epsscale{0.7}
487: %\plotone{f5.eps}
488: \includegraphics[width=0.7\textwidth,trim=0in 0.7in 0in 0in]{f5.eps}
489: \label{compltype}
490: \caption[compltype]{Completeness as a function of $R$ for source identification and redshift
491: measurement (histograms). The points show the fraction of identified BLL with measured
492: redshift; small numbers lead to substantial error bars.
493: }
494: \end{figure}
495:
496: We defer full discussion of the sample properties until we reach our expected
497: 95\% completeness to $R=23$. However, it is already interesting to examine the redshift
498: distribution of the sources detected to date (Figure 6). The non-BLL (largely
499: FSRQ) distribution peaks at
500: $z\approx 1.3$ and has an exponential fall-off (${\rm d}N/{\rm d}z \propto 10^{-0.6z}$)
501: to high redshift, extending to $z=5.5$.
502: From SED information on optically faint sources, we expect the high-redshift population
503: to increase somewhat in the complete CGRaBS sample, but it is clear that there will be
504: only a handful of radio-bright blazars at $z>4$. If any of these are detected by the LAT,
505: as expected, they will be particularly important targets for multiwavelength spectral
506: and variability studies. In fact, with only $\sim$40 sources at $z>3$, careful study
507: of these few high-redshift objects will be important for several LAT programs, e.g.,
508: extragalactic background light (EBL) studies and studies of jet evolution and interaction
509: with the CMBR.
510:
511: \begin{figure}[t]
512: \centering
513: %\plotone{f6.eps}
514: \includegraphics[width=0.7\textwidth,trim=0in 0.7in 0in 0in]{f6.eps}
515: \label{zdist}
516: \caption[zdist]{Redshift distributions for the (partly complete) CGRaBS survey. The solid-line histogram
517: shows FSRQs. The short-dashed histogram gives the redshift distribution for solved BLLs.
518: The long-dashed histogram shows a variety of other AGN (NLRGs, passive ellipticals, etc.),
519: which contribute only at very low redshift.}
520: \end{figure}
521:
522: We are also assembling an important new sample of radio-bright BLLs. To date,
523: we have 133 sources definitively classified as BLLs, but this will likely grow since
524: a substantial number of other sources have observed BLL-like spectra but need somewhat
525: improved S/N observations to exclude emission lines with EW $\ge$ 5 \AA\ throughout the observed
526: spectrum. Among the brighter sources $\sim$15\% are BLL; at this incidence, we expect
527: $\sim$245 sources to have a final BLL
528: classification. As noted, it will be very tough to obtain redshifts of the faintest
529: BLLs. However, the 70 redshifts already in hand represent a substantial radio-bright
530: sample. For example, it is twice the size of the 1 Jy sample \citep{stet91}
531: and extends to nearly twice the redshift. At present, we have 11 BLLs at $z>1$, a
532: third of all known $z>1$ BLLs, so that the full survey should be useful for probing
533: evolution of this population.
534:
535: Of course, the most important application of the CGRaBS catalog is the identification
536: with other all-sky samples and the generation of multiwavelength SEDs. We are already
537: examining the radio to X-ray spectra of these sources and eagerly look forward to
538: the upcoming sky surveys with {\it AGILE}, the air-\v{C}erenkov TeV observatories, and
539: especially {\it GLAST}, which will measure the $\gamma$-ray power peak expected for many of these
540: sources.
541:
542: \acknowledgments
543:
544: S.\ E.\ H.\ was supported by SLAC under DOE contract DE-AC03-76SF00515.
545:
546: The Hobby*Eberly Telescope (HET) is a joint project of the University of
547: Texas at Austin, the Pennsylvania State University, Stanford University,
548: Ludwig-Maximilians-Universit\"{a}t M\"{u}nchen, and
549: Georg-August-Universit\"{a}t G\"{o}ttingen. The HET is named in honor
550: of its principal benefactors, William P.\ Hobby and Robert E.\ Eberly.
551:
552: The Marcario Low Resolution Spectrograph is named for Mike Marcario of
553: High Lonesome Optics, who fabricated several optics for the instrument
554: but died before its completion. The LRS is a joint project of the
555: Hobby*Eberly Telescope partnership and the Instituto de Astronom\'{i}a
556: de la Universidad Nacional Aut\'{o}noma de M\'{e}xico.
557:
558: This paper includes data taken at the McDonald Observatory of the
559: University of Texas at Austin.
560:
561: The data in this paper are based partly on observations obtained at the
562: Hale Telescope, Palomar Observatory, as part of a collaborative
563: agreement between the California Institute of Technology, its divisions
564: Caltech Optical Observatories and the Jet Propulsion Laboratory
565: (operated for NASA), and Cornell University.
566:
567: Some of the data presented herein were obtained at the W.\ M.\ Keck
568: Observatory, which is operated as a scientific partnership among the
569: California Institute of Technology, the University of California and the
570: National Aeronautics and Space Administration. The Observatory was made
571: possible by the generous financial support of the W.\ M.\ Keck
572: Foundation.
573:
574: The authors wish to recognize and acknowledge the very significant
575: cultural role and reverence that the summit of Mauna Kea has always
576: had within the indigenous Hawaiian community. We are most fortunate
577: to have the opportunity to conduct observations from this mountain.
578:
579: Some of the data in this paper are based on observations made with ESO
580: telescopes at the La Silla Observatory under program 077.B-0056(A) and
581: ID 078.B-0275(B).
582:
583: Funding for the SDSS and SDSS-II has been provided by the Alfred P.\
584: Sloan Foundation, the Participating Institutions, the National Science
585: Foundation, the U.\ S.\ Department of Energy, the National Aeronautics
586: and Space Administration, the Japanese Monbukagakusho, the Max Planck
587: Society, and the Higher Education Funding Council for England. The SDSS
588: Web site is {\tt http://www.sdss.org/}.
589:
590: The SDSS is managed by the Astrophysical Research Consortium for the
591: Participating Institutions. The Participating Institutions are the
592: American Museum of Natural History, Astrophysical Institute Potsdam,
593: University of Basel, University of Cambridge, Case Western Reserve
594: University, University of Chicago, Drexel University, Fermilab, the
595: Institute for Advanced Study, the Japan Participation Group, Johns
596: Hopkins University, the Joint Institute for Nuclear Astrophysics, the
597: Kavli Institute for Particle Astrophysics and Cosmology, the Korean
598: Scientist Group, the Chinese Academy of Sciences (LAMOST), Los Alamos
599: National Laboratory, the Max Planck Institute for Astronomy (MPIA), the
600: Max Planck Institute for Astrophysics (MPA), New Mexico State
601: University, Ohio State University, University of Pittsburgh, University
602: of Portsmouth, Princeton University, the United States Naval
603: Observatory, and the University of Washington.
604:
605: This research has made use of the NASA/IPAC Extragalactic Database (NED), which
606: is operated by the Jet Propulsion Laboratory, California Institute of Technology,
607: under contract with the National Aeronautics and Space Administration.
608:
609:
610: \begin{thebibliography}{}
611: \bibitem[Adelman-McCarthy et al.(2007)]{sdss}Adelman-McCarthy, J.\ et al.\ 2007, ApJS, in press.
612: \bibitem[Bock et al.(1999)]{sumss}Bock, D.\ C.-J., Large, M.\ I., \& Sadler, E.\ M.\ 1999, AJ, 117, 1578.
613: \bibitem[Condon et al.(1998)]{nvss}Condon, J.\ J.\ et al.\ 1998, AJ, 115, 1693.
614: \bibitem[Hartman et al.(1993)]{hart93}Hartman, R.\ C.\ et al.\ 1993, ApJ, 407, L41.
615: \bibitem[Hartman et al.(1999)]{3eg}Hartman, R.\ C.\ et al.\ 1999, ApJS, 123, 79.
616: \bibitem[Healey et al.(2007)]{crates}Healey, S.\ E.\ et al.\ 2007, ApJS, 171, 61.
617: \bibitem[Hill et al.(1998)]{lrs}Hill, G.\ J.\ et al.\ 1998, SPIE, 3355, 375.
618: \bibitem[Kniffen et al.(1993)]{kniffen}Kniffen, D.\ A.\ et al.\ 1993, ApJ, 411, 133.
619: \bibitem[March\~{a} et al.(1996)]{marcha}March\~{a}, M.\ J.\ M.\ et al.\ 1996, MNRAS, 281, 425.
620: \bibitem[Massaro et al.(2007)]{asdc}Massaro, E.\ et al.\ 2007, {\tt http://www.asdc.asi.it/bzcat/}
621: \bibitem[Mattox et al.(2001)]{mattox}Mattox, J.\ R.\ et al.\ 2001, ApJS, 135, 155.
622: \bibitem[Monet et al.(2003)]{usno}Monet, D.\ G.\ et al.\ 2003, AJ, 125, 984.
623: \bibitem[Myers et al.(2003)]{class}Myers, S.\ T.\ et al.\ 2003, MNRAS, 341, 1.
624: \bibitem[Padovani \& Giommi(1995)]{hbl}Padovani, P.\ \& Giommi, P.\ 1995, ApJ, 444, 567.
625: \bibitem[Schneider et al.(2007)]{sdssq}Schneider, D.\ P.\ et al.\ 2007, AJ, 134, 102.
626: \bibitem[Schlegel, Finkbeiner \& Davis(1998)]{schleg}Schlegel, D.\ J., Finkbeiner, D.\ P.\ \& Davis, M.\ 1998, ApJ, 500, 525.
627: \bibitem[Shetrone et al.(2007)]{shetrone}Shetrone, M. et al.\ 2007, PASP, 119, 556.
628: \bibitem[Sowards-Emmerd et al.(2003)]{srm03}Sowards-Emmerd, D.\ et al.\ 2003, ApJ, 590, 109.
629: %\bibitem[Sowards-Emmerd et al.(2004)]{srm04}Sowards-Emmerd, D.\ et al.\ 2004, ApJ, 609, 564.
630: \bibitem[Sowards-Emmerd et al.(2005)]{srm05}Sowards-Emmerd, D.\ et al.\ 2005, ApJ, 626, 95.
631: \bibitem[Stickel et al.(1991)]{stet91}Stickel, M.\ et al.\ 1991, ApJ, 374, 431.
632: \bibitem[Turriziani et al.(2007)]{roxa}Turriziani, S.\ et al.\ 2007, astro-ph/0705.1498.
633: \bibitem[V\'{e}ron-Cetty \& V\'{e}ron(2006)]{veron}V\'{e}ron-Cetty, M.-P.\ \& V\'{e}ron, P.\ 2006, A\&A, 455, 773.
634: \bibitem[Voges et al.(1999)]{rass}Voges, W.\ et al.\ 1999, A\&A, 349, 389.
635: \end{thebibliography}
636:
637: \end{document}
638: