Discovery of a_massive_selected_galaxy_cluster

ApJ, in press
Preprint typeset using L TEX style emulateapj v. 5/2/11
A

IDCS J1426+3508: DISCOVERY OF A MASSIVE, IR-SELECTED GALAXY CLUSTER AT Z = 1.75
S. A. Stanford,1,2 M. Brodwin,3,4 Anthony H. Gonzalez,5 Greg Zeimann,1 Daniel Stern,6 Arjun Dey,7
P. R. Eisenhardt,6 Gregory F. Snyder,4 and C. Mancone5
ApJ, in press

ABSTRACT
arXiv:1205.3786v1 [astro-ph.CO] 16 May 2012

We report the discovery of an IR-selected massive galaxy cluster in the IRAC Distant Cluster Survey
(IDCS). We present new data from the Hubble Space Telescope and the W. M. Keck Observatory that
spectroscopically confirm IDCS J1426+3508 at z = 1.75. Moreover, the cluster is detected in archival
Chandra data as an extended X-ray source, comprising 54 counts after the removal of point sources.
We calculate an X-ray luminosity of L0.5−2keV = (5.5 ± 1.2) × 1044 ergs s−1 within r = 60 arcsec (∼ 1
Mpc diameter), which implies M200,Lx = (5.6 ± 1.6) × 1014 M⊙ . IDCS J1426+3508 appears to be an
exceptionally massive cluster for its redshift.
Subject headings: galaxies: clusters: individual — galaxies: distances and redshifts — galaxies: evo-
lution

1. INTRODUCTION lished by Gobat et al. (2011) appear to be less than
Galaxy clusters present the opportunity for address- 1 × 1014 M⊙ . Santos et al. (2011) and Fassbender et al.
ing two main issues in astrophysics. Cosmological pa- (2011) have published spectroscopically confirmed X-ray
rameters may be constrained with knowledge of the selected clusters at z = 1.58 and z = 1.56, respec-
abundance of clusters if the selection function, red- tively, both of which appear to have cluster masses of a
shifts, and masses are adequately known (Haiman et al. few×1014 M⊙ , although in the former case only 3 mem-
2001; Holder et al. 2001). Even individual clusters may ber galaxies have spectroscopic redshifts. Thus there are
be useful if they are sufficiently massive and at suffi- only 1–2 galaxy clusters at z > 1.5 that are massive, con-
ciently high redshift because they trace the extreme tail firmed, and published. A number of proto–clusters have
of the cosmological density field (Matarrese et al. 2000; been identified and confirmed at z > 2 (Pentericci et al.
Mortonson et al. 2011). 2000; Venemans et al. 2007; Capak et al. 2011). While
Impressive progress has been made over the past very interesting, these systems appear to be in the very
decade in finding and characterizing galaxy clusters at early stages of cluster formation–their masses remain dif-
1 < z < 1.5. At z 1.5 the number of massive, high red- ficult to estimate and their natures difficult to interpret.
shift clusters that have been identified and confirmed is The other main astrophysical use for galaxy clusters
still very limited. The accounting depends on the defini- is to help us understand galaxy evolution. In particu-
tions both of what constitutes a massive cluster and what lar, they contain the majority of the massive early-type
constitutes confirmation. Here we assume a high-redshift galaxies in the universe, so these environments offer ex-
cluster massive enough to be useful for constraining cos- cellent places in which to study such galaxies. To trace
mological parameters has an M200 mass (the mass within the evolution of massive early-type galaxies over their
the region where the cluster overdensity is 200 times the full lifetime, we must identify and study the precursor
critical density) of at least 1 × 1014 M⊙ . Such objects cluster population over a large redshift range. This kind
are the progenitors of present-day clusters with masses of of archaeology requires the evolutionary precursors to be
∼ 5 × 1014 M⊙ . We suggest that confirmation of a clus- identified in large, statistically useful samples, which are
ter candidate requires at least 5 spectroscopic redshift sensitive down to the group scale at relatively high red-
members within 2 Mpc. shift, since the massive clusters at z < 1 are built from
At z > 1.5 several systems have been published in groups and low-mass clusters at z > 1.
the literature. The estimated masses of the cluster at Identifying and then characterizing cluster samples ad-
z = 1.62 identified by both Papovich et al. (2010) and equate to these two tasks has been challenging. Op-
Tanaka et al. (2010), the cluster at z = 1.75 reported by tical methods of finding clusters tend to succeed up
Henry et al. (2010), and the cluster at z = 2.07 pub- to redshifts only slightly beyond unity. Neither X-ray
nor Sunyaev-Zel’dovich (SZ) cluster surveys currently
1 Department of Physics, University of California, One Shields have the sensitivity to reach cluster masses down to
Avenue, Davis, CA 95616
2 Institute of Geophysics and Planetary Physics, Lawrence
1 × 1014 M⊙ at z ∼ 1.5 and above over appreciable areas.
Livermore National Laboratory, Livermore, CA 94550
The IRAC Shallow Cluster Survey (ISCS,
3 Department of Physics and Astronomy, 5110 Rockhill Road Eisenhardt et al. 2008) originally was designed to
University of Missouri, Kansas City, MO, 64110 create a stellar mass-selected selected sample of galaxy
4 Harvard-Smithsonian Center for Astrophysics, 60 Garden
clusters spanning 0 < z < 2. The ISCS is drawn from
Street, Cambridge, MA 02138
5 Department of Astronomy, University of Florida,
the Spitzer/IRAC Shallow Survey (Eisenhardt et al.
Gainesville, FL 32611 2004), which imaged most of the Bo¨tes field in the
o
6 Jet Propulsion Laboratory, California Institute of Technol-
NOAO Deep Wide-Field Survey (Jannuzi & Dey 1999).
ogy, Pasadena, CA 91109 Clusters were identified by searching for 3-dimensional
7 NOAO, 950 North Cherry Avenue, Tucson, AZ 85719

2 Stanford et al.

spatial overdensities in a 4.5µm selected galaxy sample nated by light from objects). We verified this procedure
with robust photometric redshifts (Brodwin et al. 2006). by confirming that it produces the correct scaling in pho-
The selection is independent of the presence of a red tometric scatter of sources detected in the sets of dither
sequence. There are 335 clusters and groups in the ISCS images before making the final stack.
sample, identified over 7.25 deg2 within the Bo¨tes field,
o Morphologies of the galaxies in the WFC3 image were
and 1/3 of the groups/clusters are at z > 1. We have determined by using Galapagos (Haussler et al. 2011) to
spectroscopically confirmed over 20 clusters spanning run GALFIT (Peng et al. 2010). Galapagos measures
1 < z < 1.5. (Stanford et al. 2005; Brodwin et al. 2006; the sky around every galaxy and uses the basic isopho-
Elston et al. 2006; Eisenhardt et al. 2008; Brodwin et al. tal parameters measured by source extractor to generate
2011, Brodwin et al. in prep, Zeimann et al. in prep). a first guess set of Sersi¸ parameters for each galaxy.
c
We have begun extending the work of the ISCS to more It then uses these values to fit a single Sersi¸ profile to
c
effectively target the higher redshift range by making every galaxy with GALFIT, simultaneously fitting close
use of the deeper IRAC imaging obtained by the Spitzer neighbors. We use the Sersi¸ index n to classify galaxy
c
Deep, Wide Field Survey (SDWFS, Ashby et al. 2009). morphologies as being either early-type (n > 2.5) or late-
The IRAC exposure time of SDWFS is 4× that of the type (n ≤ 2.5).
IRAC data used in the ISCS, which has allowed us to bet- 3. SPECTROSCOPY
ter identify cluster candidates at z > 1.5 in the Bo¨tes
o
field as part of the new IRAC Distant Cluster Survey 3.1. Keck Optical Spectra
(IDCS). Here we report the first spectroscopic confirma- Spectra with the Low-Resolution Imaging Spectro-
tion of one of these candidates, IDCS J1426+3508 at graph (LRIS; Oke et al. 1995) on Keck I were acquired
z = 1.75, which we believe to be a massive cluster. Com- on UT 2011 April 28 and 29 using 1.1′′ × 10′′ slitlets, the
panion papers in this volume report on the mass of this G400/8500 grating on the red side, the D680 dichroic,
cluster (Brodwin et al. 2012) and on a gravitational arc and the 400/3400 grism on the blue side. On the
(Gonzalez et al. 2012). We present the optical and in- ′′
first night when the seeing was 0. 6 and conditions were
frared imaging in §2, the spectroscopic observations and mostly clear, data were obtained in one mask in six 1200 s
resulting redshifts in §3 and §4, respectively, and the X- exposures. On the second night when the seeing var-
ray observations in §5. We use Vega magnitudes and a ied in the range 0.6 − 0.8 arcsec with light cirrus, data
WMAP7+BAO+H0 ΛCDM cosmology (Komatsu et al. were obtained in a second mask for ten 1200 s expo-
2011): ΩM = 0.272, ΩΛ = 0.728, and H0 = 70.4 km s−1 sures. These data were split into slitlets which were sep-
Mpc−1 . arately reduced following standard procedures. The rela-
2. OPTICAL AND NEAR-IR IMAGING
tive spectral response was calibrated via longslit (with a
width of 1.0 arcsec) observations of Wolf 1346 and Feige
The cluster candidate was originally identified using 34 (Massey & Gronwall 1990).
the SDWFS data matched with the NDWFS optical
data following the same procedures described in detail 3.2. WFC3 Grism Spectra
in Eisenhardt et al. (2008). A significant overdensity in The WFC3 data were obtained on 2010 November 6.
the 3-dimensional space of (RA, Dec, and photometric A total exposure time of 11247 s was used for the G102
redshift) was selected for further study. A color image observation and 2011 s for G141; a single pointing was
made from the NDWFS optical + IRAC data is shown used in both cases. We targeted the cluster candidate
in the left panel of Figure 1 where a tight red group of with both IR prisms so as to provide continuous wave-
galaxies is visible. The photometric redshift estimate of length coverage of nearly 1 µm, allowing for conclusive
the cluster is z ∼ 1.8. identifications of spectral features in the redshift range
Deeper optical and NIR follow-up imaging was ob- of interest. The FWHM of an unresolved emission line is
tained with the Hubble Space Telescope using ACS and ∼2 pixels, which corresponds to a spectral resolution of
WFC3. F814W exposures were obtained with ACS in 49 ˚ for G102 and 93 ˚ for G141. The low grism resolu-
A A
one pointing for 8 × 564 s and reduced using standard tion blends some of the more common emission lines such
procedures. The WFC3 data were obtained with the as Hα+N[II] and [OIII]5007+[OIII]4959. This typically
F160W filter in two slightly overlapping pointings, each results in a redshift uncertainty of σz ∼ 0.01.
comprising 700 s of integration time in dithered expo- In slitless spectroscopy, a direct image is a necessary
sures. The WFC3 data were reduced using standard pro- companion to the grism image in order to zero-point
cedures with the MultiDrizzle software. A pseudo-color the wavelength scale and to properly extract spectra.
image constructed from the registered ACS and WFC3 We chose broadband filters that closely matched the
imaging is shown in the right panel of Figure 1. grism spectral coverage: F105W for G102 and F140W
Photometry was measured from the HST imaging us- for G141. The object positions and sizes measured from
ing SExtractor (Bertin & Arnouts 1996) in dual image the direct images are used to establish the location, wave-
mode with sources detected in the WFC3 image. Col- length zero point, and spectral extraction widths of the
ors were measured in 0.8 arcsec diameter apertures, and objects in grism images. The data were reduced using
MAG AUTO is used as a“total” magnitude. Our pho- aXe (Kummel et al. 2008). The calibration files1 used
tometric uncertainties are dominated by sky shot noise. were the best available at the time of the reductions.
We estimated the uncertainties from the distribution of The steps used to extract spectra are very similar to that
sky background measurements in 5000 randomly placed found in WFC3 Grism Cookbook2 and a more detailed
0.8′′ apertures. These sky measurements are roughly nor-
1
mally distributed; we estimate σsky by fitting a Gaussian http://www.stsci.edu/hst/wfc3/analysis/grism obs/calibrations/
2 http://www.stsci.edu/hst/wfc3/analysis/grism obscookbook.html
to the left-half of the distribution (which is uncontami-

IDCS J1426+3508 at z = 1.75 3

Fig. 1.— (left) Color image covering 3 × 3 arcmin using imaging from the NDWFS BW and I, and IRAC 4.5 µm data centered on IDCS
J1426+3508. The Q marks the quasar in the cluster. (right) Pseudo-color HST image made from the ACS/F814W and WFC3/F160W
images with the green contours illustrating the X-ray emission. The dashed red circle is centered on the quasar in the cluster, and the blue
dashed circle is centered on a non-member radio-loud AGN. The radii of these two dashed circles is 5 arcsec, the same size as was used to
mask these point sources in the X-ray analysis. In both panels the yellow boxes are spec-z confirmed members, and a 30′′ (260 kpc) scale
bar is given.
description will be presented in Zeimann et al. (in prepa- used with an SDSS LRG template spectrum3 , redshifted
ration). to z = 1.75, to create a mock grism image. A 1-D extrac-
tion was performed with the same reduction procedure
4. REDSHIFT MEASUREMENTS as was used with the actual WFC3 grism observations.
The reduced optical spectra were visually inspected to The redshifts determined from both the LRIS and WFC3
determine redshifts. Despite the long integrations on spectroscopy on the identified cluster members are sum-
a 10 m telescope using a spectrograph with new red- marized in Table 1, along with magnitudes and colors
sensitive CCDs, sufficiently good LRIS spectra were ob- obtained from the HST images. The overall redshifts ob-
tained on only two phot-z selected objects which were tained in the vicinity of the cluster candidate with the
found to have features such as D4000, B2640 and the LRIS and WFC3 grism data are presented in Figure 4.
MgIIλ2800 absorption line that are characteristic of older 5. X-RAY OBSERVATIONS
stellar populations. These features indicate z = 1.75 for
both objects, which were confirmed by the WFC3 IR The Bo¨tes field has been surveyed previously
o
grism spectroscopy. with ACIS-I onboard the Chandra X-Ray Observa-
The NIR grism spectra were first visually in- tory (Murray et al. 2005; Kenter et al. 2005; Brand et al.
spected. Emission lines were identified as be- 2006). At the position of IDCS J1426+3508, exposures
ing blended Hα+[NII], [OIII]λ5007+[OIII]λ4959, totalling 9.5 ks are available from the Chandra archive.
Hβ, or [OII]λ3727. The spectra were also cross- These data are split between an observation of 4.8 ks on
correlated with spectral templates (taken from UT 2006 July 30 (ObsID 3621) and an observation of 4.7
www.sdss.org/dr7/algorithms/spectemplates/index.html) ks on UT 2006 August 21 (ObsID 7381). We processed
to automatically determine redshifts where feasible the data following standard procedures using the Chan-
(Zeimann et al., in prep). Contamination from overlap- dra Interactive Analysis of Observations (CIAO; V4.2)
ping spectra was estimated using a Gaussian emission software. We initially identified good-time intervals for
model, scaled by the measured broadband magnitudes the exposures, yielding a total effective exposure time of
from the direct images (F105W or F140W; see the 8.3 ks for IDCS J1426+3508.
WFC3 Grism Cookbook for more details). This process The cluster is clearly detected as an extended source
is handled in the standard reduction package of aXe. in both individual exposures, as well as in the stacked
For the ETG spectra, we experimented with a range exposure. The cluster is approximately 6.5 arcmin off-
of SDSS galaxy templates and found that the LRG axis in both exposures, for which the Chandra point-
template was the best match to our grism data. source 50% encircled energy radius is 2 arcsec at 1.5
The spectra of cluster members are shown in Figure 2 keV. This complicates the X-ray analysis, as an opti-
and Figure 3. The latter shows in the bottom panel cally bright quasar confirmed to be in the cluster (see
a simulated early-type galaxy spectrum, as observed by
3 http://www.sdss.org/dr7/algorithms/spectemplates/index.html
WFC3 with our observational parameters. aXeSim was

4 Stanford et al.

1 J142634.4+350825
z = 1.75 [OII] Hβ [OIII]

0.5

0

30 J142632.4+350830
20
erg s cm Å )
−2 −1

10

0

J142628.1+350829
−1

1

0.5

0
−18
Flux (10

J142634.4+350822
3 z = 1.75 [OII] Hβ [OIII]
2
1
0
−1
J142632.8+350844
0.6 z = 1.74 [OII] Hβ [OIII]
0.4
0.2
0
−0.2
0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6
Wavelength (µm)

Fig. 2.— WFC3 spectra of the five cluster members that exhibit emission lines are plotted above. The solid black histograms are the
spectra from the G102 and G141 grisms. The dot-dashed cyan line is the estimate of contamination from overlapping spectra which is
subtracted off in the final stage of reduction. The solid red line is the 1-σ flux error. The vertical green lines which are labeled are the
detected or expected emission from the [OII]λ3727, Hβ, and [OIII]λ5007 lines at the nominal cluster redshift. The bright, power-law
spectrum second from the top is a QSO, previously identified in AGES optical spectroscopy (Kochanek et al., in preparation); in this panel
a QSO template (SDSS; Vanden Berk et al. 2001) is shown by the magenta line.

TABLE 1
Spectroscopic Cluster Members

ID R.A. (J2000) Decl. (J2000) z ∆z Instrument F160W F814W-F160W
J142632.9+350823a 14:26:32.95 35:08:23.6 1.75 0.01 WFC3/LRISb 19.25 4.47
J142632.5+350822 14:26:32.55 35:08:22.5 1.75 0.01 WFC3/LRISc 20.41 4.16
J142632.4+350830d 14:26:32.40 35:08:30.8 1.746e 0.01 WFC3 18.59 1.08
J142632.8+350844 14:26:32.85 35:08:44.4 1.74 0.01 WFC3 22.81 1.20
J142634.4+350825 14:26:34.43 35:08:25.1 1.75 0.01 WFC3 22.63 1.14
J142634.4+350822 14:26:34.47 35:08:22.4 1.75 0.01 WFC3 22.27 0.90
J142628.1+350829 14:26:28.15 35:08:29.7 1.75 0.01 WFC3 23.00 1.43

a Brightest Cluster Galaxy
b LRIS spectrum shows a break at 2640 ˚ and a very red continuum consistent with z = 1.75
A
c LRIS spectrum shows a red continuum and a MgIIλ2800 absorption feature consistent with z = 1.75
d QSO
e AGES redshift; WFC3 grism redshift is 1.77

IDCS J1426+3508 at z = 1.75 5

3
J142632.9+350823 G Hγ Hβ [OIII] Mg
z = 1.75
2 H+K

1
Å )
−1

0 0.9 0.94 0.98
−2

J142632.5+350822
cm

z = 1.75
Wavelength (µm)
G Hγ Hβ [OIII] Mg
1
−1

H+K
erg s

0.5
−18

0
Flux (10

Simulated Spectrum
3 z = 1.75

2

1

0

1 1.05 1.1 1.15 1.2 1.25 1.3 1.35 1.4 1.45
Wavelength (µm)
Fig. 3.— WFC3 spectra of the two cluster members with early-type spectra in the top two panels; the bottom panel shows a simulated
spectrum, described in the text, for reference. The solid black histograms are the spectra from the G102 and G141 grisms. The dot-dashed
blue line is the estimation of contamination from overlapping spectra which is subtracted off in the final stage of reduction. The solid red
line is the 1-σ flux error. The vertical green lines are the expected locations of the [OII]λ3727, Hβ, and [OIII]λ5007 lines. The vertical
red lines are the expected locations for the following absorption features: Ca H+K, the G-band, Hγ, and MgIIλ2800. The magenta lines
represent the SDSS LRG template fitted to the observed spectra. The inset spectrum in the top panel shows the LRIS spectrum (black
solid line) in the vicinity of the D4000 break, along with the template fit (magenta solid line), solidifying the reality of this feature seen in
the grism spectrum.

6 Stanford et al.

flux. We then extracted cluster source counts in the
0.5 - 7 keV range within a 1 arcmin radius aperture
centered on the cluster BCG. This aperture approxi-
mately corresponds to a radius of 500 kpc at the clus-
ter redshift. Response matrices and effective areas were
then determined for each detected source. Within the
measurement aperture, there are 54 ± 10 background-
corrected counts in the 0.5 − 7 keV range, after masking
out the two central AGN. We used XSPEC (V12.6.0) to
fit the background-subtracted X-ray spectrum with the
MEKAL hot, diffuse gas model (Mewe et al. 1985) us-
ing the Wisconsin photo-electric absorption cross-section
(Morrison & McCammon 1983). The temperature was
fixed at 5 keV and the abundance at 0.3 M⊙ , with a
Galactic absorption of 1.3 ×1020 cm−2 at the target po-
sition. We determined a Galactic absorption-corrected
flux of (3.1 ± 0.7) × 10−14 ergs cm−2 s−1 in the 0.5 − 2.0
keV range, which translates to an X-ray luminosity of
(5.5 ± 1.2) × 1044 ergs s−1 at z = 1.75. The X-ray flux
changes by only 7% if the X-ray temperature is varied
from 4 to 6 keV.
Using the M500 −LX relation of Vikhlinin et al. (2009),
we estimate from the luminosity that M500,LX = (3.5 ±
1.0) × 1014 M⊙ . We caution that the use of this scal-
ing relation requires a significant extrapolation in red-
shift. This and other systematic uncertainties (such
Fig. 4.— Redshift histogram resulting from the LRIS and WFC3
spectroscopic observations. The green bars show all spectroscopic as removal of X-ray point sources) are expected to
redshifts of all qualities. The blue histogram shows only the robust dominate over statistical errors. To estimate a total
redshifts (i.e., quality A or B), and the red histogram shows the cluster mass, we next assume an NFW density pro-
subset of these which lie within 30” of the brightest cluster galaxy. file (Navarro et al. 1997) and the mass-concentration re-
The inset shows a detail of the redshift histogram near the cluster
redshift; its width is consistent with the uncertainties of the grism lation of (Duffy et al. 2008). The resulting M200 =
redshifts. (5.6 ± 1.6) × 1014 M⊙ , where the uncertainty is deter-
mined only from the statistical measurement error.
Section 5) is only 9 arcsec from the brightest cluster
galaxy (BCG). In addition, we identify an X-ray point 6. GALAXY POPULATIONS
source associated with a radio source that is 12 arcsec Figure 5 presents the photometric and morphological
to the SW, at 14:26:32.2, +35:08:14.9. This source, con- information for all the objects in the 7.7 arcmin2 area
firmed as an emission line galaxy at z = 1.535 in our centered on the cluster where the ACS and WFC3 imag-
WFC3 grism data, has an integrated 21 cm flux density ing overlap. While it is possible to see a red sequence of
of 95.3 mJy from the FIRST survey (Becker et al. 1995). early-type galaxies in Figure 5, the spread and location is
Given the signal to noise ratio and the large off-axis an- different from that of the red sequence in massive clusters
gle of the available X-ray observations, it is challenging at 1 < z < 1.5. To isolate a sample of probable cluster
to disentangle the extended cluster emission from the members in the color-magnitude diagram we carry out
point source contributions. However, as seen in the right the following multi-step procedure. First, we determined
panel of Figure 1, IDCS J1426+3508 is clearly associ- the relative offset ∆ between the object colors and the
ated with diffuse X-ray emission that extends beyond expected color-magnitude relation at this redshift (based
the point-source contributions from the two AGN. We on a model for Coma with zf = 3; see Eisenhardt et al.
see no evidence in the WFC3 grism spectroscopy, which (2007) for details). The range −0.5 < ∆ < 1.0 (corre-
covers the central 2 arcmin of the cluster, for other AGN sponding to colors 2.3 < F814W − F160W < 3.8) is the
which could contribute to the X-ray emission. We have color cut used to select member galaxies on the red se-
also used the SDWFS IRAC photometry to construct a quence. We initially define the cluster red sequence as
two-color diagram to search for obscured AGN (following galaxies brighter than H ∗ (z) + 1.5 that are chosen by the
Stern et al. 2005) which might contribute to the mea- above color cut in ∆, where H ∗ (z) is passively evolved
sured X-ray flux. In addition to the QSO and the radio from the luminosity of Coma early-type galaxies. Most
source already described, there is one more object within of the objects in this color-selected red sequence do not
the X-ray measurement aperture which has IRAC colors have spectroscopic or photometric redshift information,
typical of obscured AGN. and so the initial red sequence sample may suffer from in-
To extract the X-ray counts due to the cluster, we terloper contamination. For the purpose of studying the
masked the three AGN (the QSO, the radio source, cluster red sequence, we choose to restrict consideration
and the IRAC AGN) using a conservative 5 arcsec ra- to morphologically selected early-type galaxies, keeping
dius aperture, corresponding to the 90% encircled en- objects with n > 2.5. We discard objects with colors
ergy radius at the observed off-axis-angle of the clus- more than two absolute deviations from the central red
ter. We expect that the unmasked flux from the two sequence color, where the deviation is the median of the
AGN to contribute only one photon to the measured absolute value of the ∆ of the potential red sequence ob-

IDCS J1426+3508 at z = 1.75 7

jects. Objects removed in the latter step are shown with the red sequence galaxies were formed at 5 < zf < 6
a gray dot in Figure 6. The remaining objects in the (Bruzual & Charlot 2003, 2007 version). This is in con-
red sequence sample are represented by the red points in trast to the cluster at z = 1.62 (Papovich et al. 2010;
Figure 6. Tanaka et al. 2010) in which the red sequence has a
To measure the color and scatter of this red sequence younger zf . The red sequences may have different me-
sample down to F160W = 21.86, we use the biweight dian colors because of the difference in the cluster masses,
estimates of location and scale (Beers et al. 1990). We in that IDCS J1426+3508 is more massive and so may
calculate the intrinsic scatter σint in the red sequence have formed earlier. Reddening by dust could cause the
sample by subtracting in quadrature the median color redder colors in IDCS J1426+3508, though there is no
error from the biweight scale estimate. The red hori- reason to suspect unusual amounts of dust in this cluster.
zontal lines in Figure 6 are offset a distance σint above Indeed, most of the spectroscopically confirmed members
and below the median color of the red sequence. We are fairly blue.
estimate uncertainties by performing these calculations
on 1000 bootstrap resamplings, from which we measure 8. CONCLUSIONS
the scale of the resulting color and σint distributions us- We have presented optical and NIR spectroscopy of
ing their median absolute deviation. The intrinsic scat- galaxies in IDCS J1426+3508, an IRAC-selected galaxy
ter in the observed colors of the red sequence galaxies is cluster candidate, which indicate this is a bona fide clus-
0.16 ± 0.06 mag; transformed to the rest frame U − V the ter of galaxies at z = 1.75. The available Chandra data
intrinsic scatter is 0.10 ± 0.04. The median color of the show a faint but clearly extended X-ray source at this
morphologically-selected red sequence sample is F814W location. Along with the centrally concentrated and reg-
− F160W = 3.96 ± 0.07, which is 0.29 ± 0.05 redder than ular distribution of red galaxies, the X-ray detection ar-
the expected color, derived from a simple passive evo- gues that IDCS J1426+3508 is gravitationally bound and
lution model with a zf = 3 (Bruzual & Charlot 2003, in a relatively relaxed state.
2007 version). A model with an earlier formation epoch, The discovery of a cluster at z = 1.75 with a soft X-
zf = 6, predicts a color which is closer to the observed ray flux greater than the ∼ 10−14 ergs cm−2 s−1 limit of
F814W − F160W colors of the morphologically-selected the upcoming eROSITA mission (Predehl et al. 2010),
red sequence. scheduled for launch in 2014, promises exciting results
The cluster includes a fairly large brightest cluster for that mission. However, we note that at the 28 arc-
galaxy. Profile fitting using GALFIT determined a Sersi¸ c sec average resolution expected for the eROSITA slew
index of 5.4 ± 0.1, re = 18.0 ± 2.5 kpc, and a to- survey, and even at the 15 arcsec on-axis resolution of
tal F160W magnitude = 18.5 ± 0.05 which is approxi- eROSITA, it would have been very difficult to distin-
mately 2.0 magnitudes brighter than M ∗ for a passively- guish the AGN from the cluster emission. Without the
evolving galaxy formed at zf = 3 (Bruzual & Charlot AGN masked out in our Chandra data, the derived X-
2003). From the F160W total magnitude, we determine ray flux for IDCS J1426+3508 would be ∼ 50% higher.
a rest frame MV = −24.7 assuming the zf = 3 passive As shown by e.g. Galametz et al. (2009), the fraction of
evolution model to transform to rest V-band. The size galaxy clusters that host luminous AGN increases rapidly
and luminosity of this BCG is similar to a BCG at z ∼ 0, with redshift. Such AGN will make it challenging to draw
which is remarkable since it is at z = 1.75. firm cosmological evolutionary results from the eROSITA
7. DISCUSSION cluster sample without higher resolution X-ray follow-up.
With the higher resolution Chandra data currently
IDCS J1426+3508 is a newly discovered galaxy clus- available from the archive, the point-source corrected X-
ter at z = 1.75, which places it among the few such ray luminosity L0.5−2.0keV = (5.5 ± 1.2) × 1044 ergs s−1 ,
systems currently known at 1.5 < z < 2, when mas- which implies M200,Lx ∼ 5 × 1014 M⊙ . This is a surpris-
sive clusters may be first forming. Using a combination ingly large cluster mass for this redshift and survey area,
of optical multi-object Keck spectroscopy and infrared as will be addressed in forthcoming papers.
HST/WFC3 grism spectroscopy, we have confirmed 7
cluster members in IDCS J1426+3508 within a radius
of 2 Mpc, all but one of which are within a radius of This work is based in part on observations obtained
250 kpc. The extended X-ray emission described in Sec- with the Chandra X-ray Observatory (CXO), under con-
tion 4 indicates that this cluster is gravitationally bound tract SV4-74018, A31 with the Smithsonian Astrophys-
and already has a mass greater than 1014 M⊙ . ical Observatory which operates the CXO for NASA.
The properties of the galaxies in the cluster indicate Support for this research was provided by NASA grant
that the cluster is far from settled in terms of star forma- G09-0150A. AHG acknowledges support from the Na-
tion. Most of the spectroscopic member galaxies are very tional Science Foundation through grant AST-0708490.
blue (c.f. Table 1) and show emission lines in their spec- This work is based in part on observations made with
tra indicative of on-going star formation or AGN activity. the Spitzer Space Telescope, which is operated by the
The red sequence itself has a larger amount of scatter, Jet Propulsion Laboratory, California Institute of Tech-
0.16 mag, compared to 0.08 in a similar observed color nology under a contract with NASA. Support for this
in XMMU J2235−2557, a slightly more massive cluster work was provided by NASA through an award issued
at somewhat lower redshift, z = 1.39 (Strazzullo et al. by JPL/Caltech. Support for HST programs 11663 and
2010). The intrinsic scatter in IDCS J1426+3508 is rea- 12203 were provided by NASA through a grant from
sonable for a population of galaxies whose stars formed at the Space Telescope Science Institute, which is oper-
zf ∼ 3.5 − 6. The median color of these galaxies matches ated by the Association of Universities for Research in
the prediction for simple passive evolution if the stars in Astronomy, Inc., under NASA contract NAS 5-26555.

8 Stanford et al.

F160W − H*(zobs)
−1 0 1 2

4.5
F814W−F160W

4.0
3.5
3.0
Ns > 2.5
2.5 Ns < 2.5
Model R.S.
2.0 Fitted R.S.

19 20 21 22
F160W
Fig. 5.— Color-magnitude diagram of IDCS J1426+3508, made from ACS and WFC3 imaging. The red and blue colored points denote
morphological classifications, based on the Sersi¸ index. The spectroscopically confirmed members are marked by the larger stars; only
c
two of the members are red enough to appear in this plot. The black dotted line represents the expected color of a passively evolving red
sequence of galaxies formed at zf = 3, with the slope based on observed Coma colors (Eisenhardt et al. 2007). The diagonal black dashed
line is the fit to the red sequence galaxies. The diagonal dashed blue line represents the 5 σ limit on the colors.

F160W − H*(zobs)
−1 0 1 2
1.5 1.5
1.0 1.0
0.5 0.5
0.0 0.0
∆

−0.5 −0.5
−1.0 −1.0
−1.5 −1.5
19 20 21 22
F160W
Fig. 6.— The left panel of the plot is similar to the color-magnitude diagram shown in Figure 5. The F814W − F160W colors have been
zero pointed to the color predicted by a passively evolving red sequence of galaxies formed at zf = 3, with the slope based on observed
Coma colors (Eisenhardt et al. 2007). The red filled points are the objects selected by the magnitude and color cuts described in the text
which have a Sersi¸ index n > 2.5. The grey-filled circles were removed from the red sequence selection. The solid red horizontal lines
c
show the one σint intrinsic scatter of the red sequence galaxies, and the black dashed line is the fit to the colors of these galaxies. The two
spectroscopically confirmed members that lie within the limited color range of this CMD are marked by the larger stars. The right side of
the plot is a histogram stacked in the colors.

IDCS J1426+3508 at z = 1.75 9

Some of the data presented herein were obtained at the Michael Brown for combining the NDWFS with SDWFS
W. M. Keck Observatory, which is operated as a scien- catalogs, Steve Murray and his XBo¨tes team for ob-
o
tific partnership among the California Institute of Tech- taining the Chandra data in the Bo¨tes field, and Alexey
o
nology, the University of California and the National Vihklinin for advice on the analysis of the Chandra data.
Aeronautics and Space Administration. The Observa- This paper would not have been possible without the
tory was made possible by the generous financial support efforts of the support staffs of the Keck Observatory,
of the W. M. Keck Foundation. This work makes use of Spitzer Space Telescope, Hubble Space Telescope, and
image data from the NOAO Deep Wide–Field Survey Chandra X-ray Observatory. We also thank the anony-
(NDWFS) as distributed by the NOAO Science Archive. mous referee for comments which helped to improve the
NOAO is operated by the Association of Universities for presentation of our results. Support for MB was pro-
Research in Astronomy (AURA), Inc., under a coopera- vided by the W. M. Keck Foundation. The work by
tive agreement with the National Science Foundation. SAS at LLNL was performed under the auspices of the
We thank Matt Ashby for creating the IRAC catalogs U. S. Department of Energy under Contract No. W-
for SDWFS, Buell Jannuzi for his work on the NDWFS, 7405-ENG-48.

REFERENCES
Ashby, M. L. N., et al. 2009, ApJ, 701, 428 Kenter, M., et al. 2005, ApJS, 161, 9
Becker, R., et al. 1995, ApJ, 450, 559 Komatsu, E., et al. 2011, ApJS, 192, 18
Beers, T., et al. 1990, AJ, 100, 32 Kubo, J. M., Stebbins, A., Annis, J., Dell’Antonio, I. P., Lin, H.,
Bertin, E., and Arnouts, S. 1996, A&A, 117, 393 Khiabanian, H., and Frieman, J. A. 2007, ApJ, 671, 1466
Bower, R. G., Bohringer, H., Briel, U. G., Ellis, R. S., Castander, K¨mmel, M., Walsh, J. R., Pirzkal, N., Kuntschner, H., and
u
F. J., and Couch, W. J. 1994, MNRAS, 268, 345 Pasquali, A. 2009, PASP, 121, 59
Brand, K., et al. 2006, ApJ, 641, 140 Massey, P., & Gronwall, C., 1990, ApJ, 358, 344
Brodwin, M., et al. 2006, ApJ, 651, 791 Matarrese, S., et al. 2000, ApJ, 541, 10
Brodwin, M., Gonzalez, A. H., Moustakas, L. A., Eisenhardt, Mewe, R., Gronenschild, E. H. B. M., and van den Oord, G. H. J.
P. R., Stanford, S. A., Stern, D., and Brown, M. J. I. 2007, 1985, A&AS, 62, 197
ApJ, 671, L93 Morrison, R. and McCammon, D. 1983, ApJ, 270, 119
Brodwin, M., et al. 2011, ApJ, 732, 33 Mortonson, M.J., et al. 2011, PhysRevD, 83, 23015
Bruzual, G., and Charlot, S. 2003, MNRAS, 344, 1000 Murray, S. S., et al. 2005, ApJS, 161, 1
Capak, P., et al. 2011, Nature, 470, 233 Navarro, J., et al. 1997, ApJ, 490, 493
Duffy, A.R., et al. 2008, MNRAS, 390, L64 Oke, J. B., et al. 1995, PASP, 107, 375
Eisenhardt, P. R. M., et al. 2004, ApJS, 154, 48 Papovich, C., et al. 2010, ApJ, 716, 1503
Eisenhardt, P. R. M., et al. 2007, ApJS, 169, 225 Peng, C.Y., et al. 2010, AJ, 139, 2097
Eisenhardt, P. R. M., et al. 2008, ApJ, 684, 905 Pentericci, L., et al. 2000, A&A, 361, L25
Elston, R. J., et al. 2006, ApJ, 639, 816 Predehl, P., et al. 2010, arXiv:1004.5219v2
Fassbender, R., et al. 2011, A&A, 527, L10 Santos, J. S., et al. 2011, A&A, 531, L15
Fruscione, A., et al. 2006, Proc. SPIE, 6270 Stanford, S. A., et al. 2005, ApJ, 634, L129
Galametz, A., et al. 2009, ApJ, 694, 1309 Stanford, S. A., et al. 2006, ApJ, 646, L13
Gobat, G., et al. 2011, A&A, 526, 133 Strazzullo, V., et al. 2010, A&A, 524, 17
Haiman, Z., et al. 2001, ApJ, 553, 545 Tanaka, M., Finoguenov, A., and Ueda, Y. 2010, ApJ, 716, L152
Haussler, B., et al. 2011, ASPC,442, 155 Vanden Berk, B., et al. 2001, AJ, 122, 549
Henry, J. P., et al. 2010, ApJ, 725, 615 Venemans, B., et al. 2007, ApJ, 716, 1503
Holder, G., et al. 2001, ApJ, 560, L111 Vihklinin, A., et al. 2009, ApJ, 692, 1033
Jannuzi, B. T. and Dey, A. 1999, in ASP Conf. Ser. 191 —
Photometric Redshifts and the Detection of High Redshift
Galaxies, p. 111

Discovery of a_massive_selected_galaxy_cluster

More Related Content

Similar to Discovery of a_massive_selected_galaxy_cluster (20)

More from Sérgio Sacani (20)

Recently uploaded (20)

Discovery of a_massive_selected_galaxy_cluster