Issue |
A&A
Volume 518, July-August 2010
Herschel: the first science highlights
|
|
---|---|---|
Article Number | L16 | |
Number of page(s) | 5 | |
Section | Letters | |
DOI | https://doi.org/10.1051/0004-6361/201014685 | |
Published online | 16 July 2010 |
Herschel: the first science highlights
LETTER TO THE EDITOR
First detection of the Sunyaev Zel'dovich
effect increment at
m![[*]](/icons/foot_motif.png)
M. Zemcov1,2 - M. Rex3 - T. D. Rawle3 - J. J. Bock1,2 - E. Egami3 - B. Altieri4 - A. W. Blain1 - F. Boone5,7 - C. R. Bridge1 - B. Clement6 - F. Combes7 - C. D. Dowell1,2 - M. Dessauges-Zavadsky8 - D. Fadda9 - O. Ilbert6 - R. J. Ivison10,11 - M. Jauzac6 - J.-P. Kneib6 - D. Lutz12 - R. Pelló5 - M. J. Pereira3 - P. G. Pérez-González13,3 - J. Richard14 - G. H. Rieke3 - G. Rodighiero15 - D. Schaerer8,5 - G. P. Smith16 - I. Valtchanov4 - G. L. Walth3 - P. van der Werf17 - M. W. Werner2
1 - California Institute of Technology, Pasadena, CA 91125, USA
2 - Jet Propulsion Laboratory, Pasadena, CA 91109, USA
3 - Steward Observatory, University of Arizona, 933 N. Cherry Ave,
Tucson, AZ 85721, USA
4 - Herschel Science Centre, ESAC, ESA, PO
Box 78, Villanueva de la Cañada, 28691 Madrid, Spain
5 - Laboratoire d'Astrophysique de Toulouse-Tarbes, Université de
Toulouse, CNRS, 14 Av. Edouard Belin, 31400 Toulouse, France
6 - Laboratoire d'Astrophysique de Marseille, CNRS - Université
Aix-Marseille, 38 rue Frédéric Joliot-Curie, 13388 Marseille Cedex 13,
France
7 - Observatoire de Paris, LERMA, 61 Av. de l'Observatoire, 75014
Paris, France
8 - Geneva Observatory, University of Geneva, 51, Ch. des Maillettes,
1290 Versoix, Switzerland
9 - NASA Herschel Science Center, California
Institute of Technology, MS 100-22, Pasadena,
CA 91125, USA
10 - UK Astronomy Technology Centre, Science and Technology Facilities
Council, Royal Observatory, Blackford Hill, Edinburgh EH9 3HJ, UK
11 - Institute for Astronomy, University of Edinburgh, Blackford Hill,
Edinburgh EH9 3HJ, UK
12 - Max-Planck-Institut für extraterrestrische Physik, Postfach 1312,
85741 Garching, Germany
13 - Departamento de Astrofísica, Facultad de CC. Físicas, Universidad
Complutense de Madrid, 28040 Madrid, Spain
14 - Institute for Computational Cosmology, Department of Physics,
Durham University, South Road, Durham DH1 3LE, UK
15 - Department of Astronomy, University of Padova, Vicolo
dell'Osservatorio 3, 35122 Padova, Italy
16 - School of Physics and Astronomy, University of Birmingham,
Edgbaston, Birmingham, B15 2TT, UK
17 - Sterrewacht Leiden, Leiden University,
PO Box 9513, 2300 RA Leiden, The Netherlands
Received 31 March 2010 / Accepted 10 May
2010
Abstract
The Sunyaev-Zel'dovich (SZ) effect is a spectral distortion of the
cosmic microwave background as observed through the hot plasma in
galaxy clusters. This distortion is a decrement in the CMB intensity
for mm, an increment at shorter
wavelengths, and small again by
m.
As part of the Herschel Lensing Survey (HLS) we
have mapped 1E0657-56 (the Bullet cluster) with SPIRE with bands
centered at 250, 350 and 500
m and have detected the SZ effect at the two
longest wavelengths. The measured SZ effect increment central
intensities are
MJy sr-1
at 350
m
and
MJy sr-1
at 500
m,
consistent with the SZ effect spectrum derived from previous
measurements at 2 mm. No other diffuse emission is detected.
The presence of the finite temperature SZ effect correction is
preferred by the SPIRE data at a significance of 2.1
,
opening the possibility that the relativistic SZ effect
correction can be constrained by SPIRE in a sample of clusters. The
results presented here have important ramifications for both sub-mm
measurements of galaxy clusters and blank field surveys with SPIRE.
Key words: cosmic background radiation -
galaxies: clusters: individual: 1E0657-56
1 Introduction
The Sunyaev-Zel'dovich (SZ) effect is a distortion of the spectral
shape of the cosmic microwave background (CMB) due to inverse Compton
scattering in the ubiquitous, hot (
K)
intracluster medium (ICM) of galaxy clusters (Sunyaev & Zeldovich 1972).
The
canonical thermal SZ spectrum is a decrement in the brightness of the
CMB as measured through galaxy clusters at mm wavelengths and an
increment at sub-mm wavelengths which passes though a null at
mm.
To correctly describe the SZ spectral distortion
when relativistic electrons are present or the cluster is moving with
respect to the CMB additional correction terms, usually termed
``relativistic'' (or ``finite temperature'') corrections and
``kinetic'' SZ effect, are required. Measurement of these corrections
is only possible using observations at multiple wavelengths, and is
expedited by measurement at wavelengths where the expected
modifications to the thermal SZ effect are largest. In the
case of
the finite temperature corrections, the largest changes expected in
the SZ increment are at wavelengths shorter than 1
m.
Decrements in emission are rare astrophysically and can be
ascribed to
the SZ effect with little ambiguity; this has lead to measurements of
the SZ effect at mm becoming
almost routine. In
comparison, measurements of the SZ effect increment are complicated by
the presence of the dusty, high redshift galaxies which constitute the
sub-mm cosmic background. Additionally, the individual sources
comprising the sub-mm background are gravitationally lensed by galaxy
clusters, the effect of which is to preferentially correlate increases
in sub-mm emission with clustering. This correlation makes
unambiguous detection of the SZ increment difficult, though successful
measurements do exist (Lamarre
et al. 1998; Komatsu
et al. 1999;
Zemcov et al. 2003;
Zemcov et al. 2007;
Nord et al. 2009).
Moreover, the presence of the sub-mm background may contaminate
measurements of the SZ effect for
mm in less
massive
galaxy clusters (Aghanim
et al. 2005). A better understanding of the
sub-mm emission associated with galaxy clusters is required.
![]() |
Figure 1:
Source subtracted maps at 250, 350 and 500 |
Open with DEXTER |
Heretofore, systematic far infrared (FIR) surveys of many galaxy clusters to large radii have been technically challenging so a complete census of sub-mm emission from clusters has been difficult to obtain. The advent of SPIRE (Griffin et al. 2010) on Herschel (Pilbratt et al. 2010) has, for the first time, provided the capability to make deep maps of clusters to large angles on the sky and to use colour information to separate the different sources of sub-mm emission present in galaxy clusters. In addition to gravitationally lensed background sources (Rex et al. 2010), emission in clusters above the confused sub-mm background may also comprise emission from galaxies in the cluster itself (Rawle et al. 2010), as well as truly diffuse emission from the SZ effect and possibly even cold dust in the ICM. SPIRE's ability to separate sources of emission based both on spatial and spectral information allows the demographics of the sub-mm emission to be measured.
In this paper, we use deep SPIRE maps of the z
= 0.3 Bullet cluster
(1E0657-56) taken as part of the Herschel Lensing
Survey (HLS,
P.I. Egami) at 250, 350, and 500 m with 18,
25 and
36 arcsec resolution to measure the SZ effect and constrain
other
diffuse emission associated with the cluster.
2 Analysis and results
The HLS is a programme to observe 40 massive clusters in the
range
0.1 < z
< 1.0; during Herschel's science
demonstration
phase 1E0657-56 was observed to full depth. The specific observation
strategy, low level data reduction and map making processes are
summarized in Egami et al.
(2010); here we begin with the calibrated
flux and noise maps at 250, 350 and 500 m (Fig. 1 of
Egami et al. 2010).
Measurement of dim, extended emission is complicated by the
presence
of background sub-mm galaxies the confusion noise from which, at
SPIRE's resolution, is measured to be 6 mJy in all three
bands for blank fields (Nguyen
et al. 2010). Indeed, in these maps we
measure an rms noise of
mJy/beam at
250, 350, 500
m,
consistent with the combination of confusion noise and
instrument noise expected from the integration time. Obviously, care
must be taken to disentangle extended emission from confused sources.
To do this, we employ the approach discussed below, the fundamental
tenet of which is to account for and remove all 250
m emission
from the two longer wavelength channels.
To generate a 250 m source catalog, sources are detected in
the 250
m
map using the StarFinder algorithm
(Diolaiti et al. 2000).
The detection threshold is set to 2.5
and is run iteratively using
SPIRE's 250
m
point response
function (PRF) to account for blending; as the purpose here is not to
generate a catalog of real sources but rather to remove all potential
sources of emission resolved in the shortest wavelength band, an
aggressive source detection level is desirable. Because the confusion
noise in these maps is already a factor of
3 above the
instrument noise level, lowering the source detection threshold does
not substantially increase the number of identified sources. The
250
m
selected sources are then subtracted from the 250
m
map to produce the source-subtracted map shown in Fig. 1.
In order to remove 250
m emission that is correlated with
350 and 500
m
emission we employ a very conservative
approach. The first step in this process is to use the 250
m
source candidate catalog to determine the possible positions of
counterparts in the other two channels. The algorithm described below
is performed independently on both the 350 and 500
m maps;
no direct comparison between the catalogs for the two longer
wavelengths is performed or necessary. For each 250
m
candidate source position, the relevant PRF is fit to the position in
the longer wavelength channel. In the fit, the position, flux and
width of the source are allowed to vary. The resulting list of long
wavelength candidate source counterparts is then compared to the
250
m
candidate source catalog. All those fits where the best fit
flux is different by >50% at 350
m or >90% at 500
m, whose
best fit width differs from the nominal PRF by >20%,
or whose best fitting position is different by a total distance of
2 map pixels or more are deemed ``unsuccessful''. The catalog
consisting of the successful fits is then subtracted from the long
wavelength map to produce a source subtracted map at that wavelength.
Table 1:
Best fitting and expected SZ effect parameters at 350 and
500 m.
Since the angular resolution of the two longer wavelength
channels is
significantly larger than that at 250 m, it is common to have
a large fraction of the fits be unsuccessful. As this is due to the
effects of confusion rather than an actual lack of a long wavelength
counterpart source, to be conservative we need to account for these
unsubtracted sources; this process is performed in two steps. First,
we generate a 250
m
map which is the sum of the 250
m
source subtracted map (shown in Fig. 1) and all
those 250
m
selected sources for which the longer wavelength
fit was unsuccessful. This corresponds to the 250
m map which
would arise if all those sources detected and removed at the longer
wavelength were also removed from the 250
m input map.
This
map is then convolved up to the longer wavelength's resolution using
the relevant PRF and rebinned to the longer wavelength map's
resolution. This map - which is equivalent to the long wavelength
source free 250
m
counterpart map - is then pixel-wise fit to
the long wavelength map to determine the scaling between them. Bright
sources tend to make such scalings difficult in SPIRE data because
such sources have individual colours which make pixel-wise colour
comparisons have large variance. However, by construction such bright
sources have been removed from the comparison so the scaling between
the colours is well described by a linear model with coefficients of
0.76 at 350
m
and 0.40 at 500
m.
The scaled 250
m
bright source removed map is then subtracted from the equivalent long
wavelength map. This has the effect of subtracting out both the
m sources
undetected in the longer wavelength map and the component of the
confused background present in the 250
m map.
![]() |
Figure 2:
Radial averages of the source subtracted and masked maps shown in
Fig. 1.
As the absolute mean of the maps are not measured by SPIRE, the points
for each band are scattered
about 0 MJy sr-1. Also plotted
are the best fitting
isothermal |
Open with DEXTER |
As a further step, the 250 m sources whose long wavelength
fits were unsuccessful are binned into a map using their 250
m
flux and position, but the PRF of the longer wavelength channel. This
undetected source model map is then threshold cut; all pixels with
values greater than the threshold are deemed possible contaminants to
the longer wavelength maps and a pixel mask is generated based on the
positions of the cut sources. The center and right most panels of
Fig. 1
show the resulting 250
m source
removed and masked maps for the 350 and 500
m channels.
The 500 m
map shown in Fig. 1
features an
extended source coincident with the measured center of the SZ effect
in this cluster Halverson
et al. (2009). If this signal is due to
extended, roughly azimuthally symmetric emission as expected from the
SZ effect then taking radial averages of the masked maps will enhance
the signal to noise ratio of this detection. As the signal to noise
ratio per map pixel is small, here we assume azimuthal symmetry and
average the maps in simple radial bins of constant 18 arcsec width;
the results of this are shown in Fig. 2.
Interestingly, both the 350 and 500
m data exhibit extended
emission at the center of the cluster; to constrain this, we fit these
data to the best fitting isothermal
model from
Halverson et al. (2009)
which has a characteristic angular size of a few
arcminutes. The resulting fits are plotted in Fig. 2;
Table 1
lists the best fitting central SZ increment
brightnesses at 350 and 500
m. For the
Halverson et al. (2009)
model parameters the probability to exceed the
measured
for the fit at 350
m
is 0.21; for 500
m
it is 0.25. The isothermal
model is a poor
statistical description of the radial average at 250
m. In
addition to the statistical uncertainties associated with the errors
on the data output from the fit, the uncertainties in the
Halverson et al. (2009)
model contribute a further 1.2 MJy sr-1at
350
m
and 1.9 MJy sr-1 at
500
m.
We have
not included the contribution of these in the uncertainties listed in
Table 1.
In order to compare the SZ effect increments measured here
with the
signal expected from previous measurements, we compile the results of
Andreani et al. (1999),
Halverson et al. (2009)
and Plagge et al. (2010).
As all of these measurements were performed with instruments working
at 2 mm, they can be averaged to create a best estimate for
the
central Comptonization parameter using the X-ray determined cluster
temperature of keV.
Including relativistic
corrections, the uncertainty weighted average y0
is
.
Table 1
lists the expected
brightness of the SZ distortion in the SPIRE bands including the
effect of the actual filter bandpasses; Fig. 3
shows the spectral distortion of the CMB brightness for these y0and
values with the SPIRE bandpasses for reference.
Good agreement is found between the expected results given the
fiducial y0 and
and the SPIRE results at both
350 and 500
m.
The relativistic SZ effect corrections required in the
presence of a
relativistic electron population have the largest effect at both the
peaks of the SZ effect and the very shortest wavelengths.
Nozawa et al. (2000)
has calculated analytic fitting formulae for these
corrections for m;
in the calculations
above we use these formulae to correct the purely thermal
SZ effect
spectrum for the presence of relativistic electrons. To determine the
change in the SPIRE result if these finite temperature corrections
were absent, we also compute the SZ spectrum without the
relativistic
electron correction and central Comptonization parameter which leaves
the brightness at 2 mm constant,
.
Figure 3
shows this pure thermal SZ effect
spectrum and Table 1
lists the corresponding 350
and 500
m
SPIRE band weighted intensities.
![]() |
Figure 3:
The SZ effect spectrum in 1E0657-56. The 2 mm uncertainty
weighted average of the measurements of Andreani
et al. (1999), Halverson
et al. (2009) and Plagge
et al. (2010) (purple asterisk) leads to the
SZ spectrum shown (solid black curve). The best fitting
SPIREmeasurements at 350 and 500 |
Open with DEXTER |
Table 1
also lists the ratio of the 350 to 500 m SZ effect intensities for the SPIRE measurement
presented here,
the finite temperature SZ effect corrected spectrum, and the pure
thermal SZ spectrum. The SPIRE and relativistic SZ effect corrected
spectrum are consistent at the 1
level, while the SPIRE
measurement and the purely thermal SZ spectrum are different by 2.1
.
This is the strongest evidence for the presence of the
relativistic corrections to date, though not a detection of their
effect on the SZ spectrum.
3 Discussion and conclusions
The possibility of detecting the SZ effect significantly shortward of its positive peak is a testament to the extraordinary capabilities of SPIRE. However, works like Zemcov et al. (2007) show that in large galaxy cluster survey samples, significant contamination to the SZ signal from bright, gravitationally lensed background sources is common. Though the properties of this particular cluster have not precluded measurement of the SZ effect - 1E0657-56 has fewer bright, lensed sources close to its SZ effect center than typical clusters which have been observed in the sub-mm and is relatively bright and broad in the SZ effect - based on this single example it is difficult to determine whether gravitational lensing of the sub-mm background will make measurements similar to this one more challenging in more typical clusters at these and other wavelengths. Data from surveys like the HLS will allow us to understand whether typical clusters are suitable for SZ effect increment detection, and how the lensed sub-mm background will affect measurements at other wavelengths.
The possibility of diffuse emission from cold dust in the ICM
has been
discussed in the past (e.g. Stickel
et al. 2002). It is
expected that, due to sputtering by energetic photons, such dust would
have a very short lifetime in the ICM environment (Draine & Salpeter 1979).
Based on the radial averages of the 250 msource subtracted
maps, where for ICM dust with reasonable temperatures the brightest
thermal emission would occur, we find no evidence for this type of
diffuse emission in this cluster. Because in any reasonable scenario
such dust emission would be faint in a z >
0.1 cluster, we expect
that targeted searches of local clusters will be more successful for
this science, though the HLS and similar surveys can provide useful
constraints.
Given the statistical uncertainties, the 350 m
SZ effect
increment measured here is slightly less than 2
larger than
would be expected from the 2 mm measurements. This may point
to
residual problems with the source subtraction procedure. More data
will allow tuning of the confused sub-mm background removal algorithm
and checks on whether biases arising from poor background removal are
endemic in a large cluster sample.
Given the presence of the SZ effect in clusters, it seems that
care
must be taken with photometry of 500 m sources within an
arcminute or so of the cluster center; such sources will be positioned
on a diffuse background so their fluxes will be biased by a small
amount (Rex et al. 2010).
As the peak of the FIR emission of dusty sources is redshifted
to
progressively longer wavelengths, high redshift sources are expected
to have exceptionally red spectral energy distributions. The results
of this work show that care must be taken when searching for sources
based on their presence in the 500 m band alone; galaxy
clusters whose SZ effects are relatively bright and compact
could well
masquerade as such sources. Due to the well known problem of
determining counterparts to sub-mm sources at other wavelengths, such
SZ effect contamination may not be immediately obvious, particularly
as cluster fields are crowded and many possible counterparts may be
present. As a corollary, searching for very red sources in confusion
limited SPIRE surveys may turn up compact clusters based on the
presence of strong 500
m emission. The search for such
extremely red SPIRE sources in the HLS and other programmes is
underway now.
This work is based in part on observations made with Herschel, a European Space Agency Cornerstone Mission with significant participation by NASA. Support for this work was provided by NASA through an award issued by JPL/Caltech.
References
- Aghanim, N., Hansen, S. H., & Lagache, G. 2005, A&A, 439, 901 Andreani, P., Böhringer, H., dall'Oglio, G., et al. 1999, ApJ, 513, 23 [NASA ADS] [CrossRef] [Google Scholar]
- Diolaiti, E., Bendinelli, O., Bonaccini, D., et al. 2000, A&AS, 147, 335 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
- Draine, B. T., & Salpeter, E. E. 1979, ApJ, 231, 77 [NASA ADS] [CrossRef] [Google Scholar]
- Egami, E., et al. 2010, A&A, 518, L12 [CrossRef] [EDP Sciences] [Google Scholar]
- Griffin, M. J., et al. 2010, A&A, 518, L3 [Google Scholar]
- Halverson, N. W., Lanting, T., Ade, P. A. R., et al. 2009, ApJ, 701, 42 [NASA ADS] [CrossRef] [Google Scholar]
- Komatsu, E., Kitayama, T., Suto, Y., et al. 1999, ApJ, 516, L1 [NASA ADS] [CrossRef] [Google Scholar]
- Lamarre, J. M., Giard, M., Pointecouteau, E., et al.1998, ApJ, 507, L5 [Google Scholar]
- Nguyen, H. T., et al. 2010, A&A, 518, L5 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
- Nord, M., Basu, K., Pacaud, F., et al. 2009, A&A, 506, 623 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
- Nozawa, S., Itoh, N., Kawana, Y., & Kohyama, Y. 2000, ApJ, 536, 31 [NASA ADS] [CrossRef] [Google Scholar]
- Pilbratt, G. L., et al. 2010, A&A, 518, L1 [CrossRef] [EDP Sciences] [Google Scholar]
- Plagge, T., Benson, B. A., Ade, P. A. R., et al. 2010, ApJ, 716, 1118 [NASA ADS] [CrossRef] [Google Scholar]
- Rawle, T. D., et al. 2010, A&A, 518, L14 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
- Rex, M., et al. 2010, A&A, 518, L13 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
- Stickel, M., Klaas, U., Lemke, D., & Mattila, K. 2002, A&A, 383, 367 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
- Sunyaev, R. A., & Zeldovich, Y. B. 1972, Comments on Astrophysics and Space Physics, 4, 173 [Google Scholar]
- Zemcov, M., Borys, C., et al. 2007, MNRAS, 376, 1073 [NASA ADS] [CrossRef] [Google Scholar]
- Zemcov, M., Halpern, M., Borys, C., et al. 2003, MNRAS, 346, 1179 [NASA ADS] [CrossRef] [Google Scholar]
Footnotes
- ...
m
- Herschel is an ESA space observatory with science instruments provided by European-led Principal Investigator consortia and with important participation from NASA. It is open for proposals for observing time from the worldwide astronomical community. Data presented in this paper were analyzed using ``The Herschel interactive processing environment (HIPE)'', a joint development by the Herschel Science Ground Segment Consortium, consisting of ESA, the NASA Herschel Science Center, and the HIFI, PACS and SPIRE consortia.
All Tables
Table 1:
Best fitting and expected SZ effect parameters at 350 and
500 m.
All Figures
![]() |
Figure 1:
Source subtracted maps at 250, 350 and 500 |
Open with DEXTER | |
In the text |
![]() |
Figure 2:
Radial averages of the source subtracted and masked maps shown in
Fig. 1.
As the absolute mean of the maps are not measured by SPIRE, the points
for each band are scattered
about 0 MJy sr-1. Also plotted
are the best fitting
isothermal |
Open with DEXTER | |
In the text |
![]() |
Figure 3:
The SZ effect spectrum in 1E0657-56. The 2 mm uncertainty
weighted average of the measurements of Andreani
et al. (1999), Halverson
et al. (2009) and Plagge
et al. (2010) (purple asterisk) leads to the
SZ spectrum shown (solid black curve). The best fitting
SPIREmeasurements at 350 and 500 |
Open with DEXTER | |
In the text |
Copyright ESO 2010
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.