The Chandra Multi-wavelength Project (ChaMP)

A Serendipitous X-ray Survey using Chandra Archival Data

The large fields-of-view (FoVs) typical of X-ray imaging instruments have long been exploited for serendipitous surveys, resulting in far-reaching and fundamental advances in our knowledge of the X-ray universe and indeed the universe as a whole. The unprecedented spatial resolution of Chandra ( ${_<\atop^{\sim}}$0.5 $^{\prime\prime}$ at field center) combined with its very low background, allows a view of the X-ray sky up to 20 times deeper than previous X-ray surveys as well as highly efficient source identification based on the accurate X-ray positions¹. A multi-institution collaboration (including: CXC, MIT, CfA, MMTO, KPNO, CTIO, UofA, SFSU, MSU, SDSS, OSU) is carrying out a serendipitous survey, the Chandra Multi-wavelength Project (ChaMP), using Chandra archival, full FoV imaging data with a wide range of exposure times. We estimate that the ChaMP X-ray sample will total $\sim\,3000$ sources in $\sim\,8$sq. degs. per year and run for 5 years, thus including statistically robust samples of rare source types such as BL Lac objects, quiescent X-ray binaries, and high-redshift clusters.

The aim of the ChaMP survey is to make use of the parts of the images which are not the primary science of the PI project. For instance, we omit fields which have a large galaxy in them or an X-ray bright cluster of galaxies. We are concentrating on fields which have, for example, a point source as the main science objective. In this way, we will assemble our sky sample from data not expected to be intensively studied by the PI's. 14% of the Chandra AO1 fields met the criteria for inclusion in ChaMP. The analysis the team will do on such topics as testing the source detection algorithms, understanding the background, and the relation between the two on a large sky area will be a result from Champ which will benefit the whole Chandra community.

To unambiguously classify the X-ray sources into the various source types - stars, galaxies, clusters, and active galactic nuclei (AGN) - optical fluxes are required. We have designed an optical imaging program to reach optical magnitude limits (20-25) matched to the X-ray flux limit for each field. The combination of three-color optical and X-ray fluxes will allow classification of > 75% of the X-ray sources both within and outside the Galactic Plane. Optical spectroscopy to determine redshifts and verify classification methods will be obtained for a significant, representative subset.

The optically identified ChaMP X-ray selected sample will enable us to address several key questions already posed by a variety of research areas in astronomy while no doubt providing the surprises that usually result from new levels of sensitivity and resolution. The scientific motivation includes:

$\bullet$ Complement the small-area, deep Chandra surveys, which will resolve the majority of the soft (0.5-2.5 keV) and hard (2-10 keV) Cosmic X-ray background (CXRB), by covering a larger area to brighter flux limits, providing strong constraints on the ${\rm log}\,N-\,{\rm log}\,S$ intermediate between pre-Chandra and Chandra deep surveys.

$\bullet$ Determine the X-ray luminosity function (LF) of AGN and other source types as a function of redshift to fainter flux levels than previously.

$\bullet$ Provide a uniform sample of AGN relatively unaffected by the biases due to line-of-sight absorption that are present in existing optical and soft-X-ray surveys. Study AGN Spectral Energy Distributions (SEDs) as a function of redshift, luminosity and AGN type to probe their primary and secondary energy sources.

$\bullet$ Identify high redshift (z$\,>\,0.5$) clusters of galaxies to constrain cosmological parameters and to determine cluster evolution by comparison with lower redshift samples.

$\bullet$ Double the number of known galactic plane cataclysmic variables (CVs), with the most luminous seen at > 10 times the current distance limit. Measure the X-ray luminosity function of accreting binary stars such as CVs and low mass X-ray binaries, (LMXBs).

$\bullet$ Study the coronal X-ray emission of main sequence stars from the onset of dynamo activity in A stars to late-M stars, where the cores are fully convective.

Chandra X-ray Surveys: entering a new age!

Previous surveys (Einstein - Stocke et al. 1991, and ROSAT - Appenzeller et al. 1998) showed that both X-ray fluxes and optical colors are needed to cleanly segregate classes of objects (e.g., distinguish stars from AGN). Large ( $\sim\,20\,-\,60$ $^{\prime\prime}$) X-ray positional uncertainties led to multiple potential optical counterparts at faint flux limits, and required costly allocations of ground-based telescope time for source identification with uncertainties often remaining. By contrast, Chandra's ${_<\atop^{\sim}}$1 $^{\prime\prime}$ astrometry and tight point spread function (PSF, $\sim\,1$ $^{\prime\prime}$ half power diameter on axis, $\sim\,10$ $^{\prime\prime}$ even 8$^\prime$ off-axis) enable unambiguous identification of the optical counterparts in images well-matched in depth. The ChaMP will be up to 20 times deeper (2 $\times 10^{-15}$ erg cm^-2 s^-1 < F _{2-10 keV} <2 $\times 10^{-13}$ erg cm^-2 s^-1) than current wide-area surveys, include soft and hard bands simultaneously (0.2-10 keV) and cover a much larger area ($\sim\,8$ deg² /year) than Chandra PI deep surveys ($\sim\,0.4$ deg²). Thus it promises rapid and profoundly new science return on several key questions at the current frontiers of astronomy.

  

Figure 13: 30 ksec Chandra and ROSAT X-ray images of field of $z\sim 0.5$cluster RXJ003033.2+261819. On the left is an actual Chandra ACIS image (chip S3, one of 6 chips, each 8$^\prime$ on a side) restricted to a soft band for ease of comparison (0.5-2.4 keV) and boxcar smoothed by 5 for clarity. The 30 ksec ROSAT PSPC image of the same field is shown at the same scale and orientation (right).

LINK TO POSTSCRIPT FILE for Figure 13

The advantages of Chandra are graphically demonstrated in Fig. 13, which shows a Chandra ACIS-S image of the z$\sim\,0.5$ cluster RXJ003033.2+261819, taken during Orbital Activation and Calibration, as compared with the ROSAT PSPC image of the same area of sky. The cluster, while barely resolved with ROSAT, is clearly an extended source in the Chandra image and the order of magnitude improvement in the X-ray positions is also clear.

The analysis of serendipitous sources in this field, shows 24 detected point sources outside the immediate cluster region (Cappi et al. 2000). A typical source has 20 counts, or $\sim\,2.4\,\times\,10^{-15}$erg cm^-2 s^-1, ROSAT detected 11 sources to 10^-14erg cm^-2 s^-1. In an $\sim\,8'\,\times\,8'\,$area surrounding the cluster, the 0.5-2 keV surface density of point sources ( $\sim\,900\,-\,1200\,{\rm deg}^{-2}$ above $1.5 \times 10^{-15}$ erg cm^-2 s^-1) exceeds by a factor $\sim 2$ that expected based on the ROSAT ${\rm log}\,N-\,{\rm log}\,S$. This high density is seen despite the relatively high Galactic absorbing column for the field ( ${\rm N_H} = 3.9\,\times\,10^{20}\,{\rm cm}^{-2}$). A non-cluster comparison field shows no such excess. Optical identifications and redshifts are needed to confirm the implied association with the cluster.

  

Figure 14: The 74 selected Cycle 1 ChaMP high-latitude fields (filled circles) are shown in RA/DEC degrees, with (0,0) as the plot center, to show the North/South distribution. Open circles are the 15 Galactic Plane fields. Circle sizes indicate Chandra exposure times ranging from 1 to 190 ksec. Larger circles correspond to longer expsure times.

LINK TO POSTSCRIPT FILE for Figure 14

ChaMP X-ray Sample

Archival Chandra Field Selection, Lead: Nancy Evans

For each Chandra observing cycle, fields will be selected which fulfill well-defined ChaMP selection criteria. Two surveys using different Galactic latitude cuts are defined. The high galactic latitude survey includes all eligible fields with $\vert b\vert>20^{\circ}$, so on average extinction E $_{B-V}\,<\,0.1$ and $N_{Hgal}\,<\,8\,\times\,10^{20}$cm^-2, facilitating comparison with previous soft X-ray observations. A smaller Galactic Plane survey (ChaMPlane, Lead: Josh Grindlay) selects $\vert b\vert\,<10^{\circ}$and exposure time $>\,30$ ksecs. We use archival, full field, imaging data only and exclude any targeted survey fields and fields containing bright optical or X-ray sources covering $>\,10\%$ of the Chandra image. The resulting ChaMP Cycle 1 field list includes 74 high Galactic latitude and 15 Galactic Plane fields (Figure 14) and is available on our web site (hea-www.harvard.edu/CHAMP/) A similar number of fields are expected each year for the ChaMP's projected 5 year duration. The PIs of each Chandra field are notified to discuss options for various levels of cooperation/collaboration.

X-ray Data Analysis

While existing Chandra data processing pipelines include a first cut at X-ray source detection, we are using the first few deep ChaMP fields to develop a semi-automated analysis procedure to include powerful tools, such as wavelet source detection which facilitates the detection of extended sources, available in CIAO (Chandra Interactive Analysis of Observations) software package². These procedures will guarantee that the ChaMP fields are processed in a uniform and efficient manner, producing a catalog with well-defined selection criteria that lists all types of X-ray sources and their X-ray properties.

Optical Follow-up, Lead: Paul Green

An essential part of any X-ray survey is the identification and classification of the X-ray sources. We have designed an optical identification program complete to log f_x/fopt=0.5, a limit which identifies >75% of the AGN for the high latitude fields (Stocke et al. 1991), and automatically includes larger fractions of any other known source types. Applying this to the Cycle 1 ChaMP field list yields limits equivalent to a V magnitude range of 20 to 25 with a mode of 24th mag.

  

Figure 15: Histogram of optical magnitude depths required to match Chandra exposures for fields selected for the ChaMP.

LINK TO POSTSCRIPT FILE for Figure 15

The wide availability of optical catalogs and surveys at the bright end of this range allows us to design an efficient identification program using existing material where possible. At fainter optical magnitudes no catalogs are currently available³. The limiting optical V magnitudes required for the ChaMP Cycle 1 list of 74 fields are in the range 20.5 ${_<\atop^{\sim}}$V ${_<\atop^{\sim}}$24.7 (Fig 15). Optical imaging will require $\sim\,36$ nights per year on a combination of 4 m and 0.9 m telescopes at NOAO (KPNO and CTIO) and 1.5m at SAO (Smithsonian Astrophysical Observatory) using the large area (4 shooter or MOSAIC) detectors. 28 nights were awarded by NOAO and SAO between Dec 1999 and Jun 2000 with 17 completed to date. We have submitted proposals to the NOAO Survey and SAO longterm programs (for the fields with brighter flux limits) to obtain the remainder of the nights assuming a similar area will be covered each year.

Optical and X-ray source properties will be combined in order to classify the sources. At high galactic latitude, SDSS $g^{\prime}, r^{\prime},$ and $i^{\prime}$ filters provide the best option for efficient source classification. Fig. 16 illustrates how 3 SDSS filters and X-ray fluxes yield a multicolor space of unprecedented AGN selection efficiency across all redshifts. All AGN are separated from stars following the the color-plane methodology of Newberg et al. (1998). Often, further discrimination will be available from optical and X-ray morphology and from X-ray spectral information.

ChaMP Archive

All ChaMP software and procedures will be documented on the ChaMP WWW site

(hea-www.harvard.edu/CHAMP/) along with the field lists for each Chandra Cycle as they become available. The ChaMP X-ray and optical data and X-ray catalog products will be made publicly available on a field-by-field basis within a year of completion of the data reduction of the full dataset for that field. We are confident that the multi-wavelength dataset so provided will prove a valuable scientific resource for the community.

- Belinda Wilkes, Paul Green

References

Appenzeller, I. et al. 1998, ApJS, 117, 319

Cappi, M. et al., 2000, ApJ, submitted

Fan, X. 1999, AJ, 117,2528

Miyaji, T., Hasinger, G., Schmidt, M. 2000, AA, in press

Newberg, H., J., Richards, G.T., Fan, X. & Laurent-Muehleison 1998, BAAS, 30, 1412

Stocke, J. T. et al 1991, ApJS, 76, 813

¹The degraded inflight energy resolution of part of the ACIS instrument does not affect the ChaMP survey limits.

²available on the CXC website asc.harvard.edu

³Once the Sloan Digital Sky Survey (SDSS) archive for all northern fields opens ( $\sim\,2002$), it will cover ChaMP imaging needs to V<23 with the same filters as the remainder of the ChaMP. While these magnitudes will not be temporally as close to Chandra observations as our own imaging, they will reduce our need for telescope time for northern fields, while still ensuring color uniformity with the ChaMP.

  

Figure 16: We have supplemented a catalog of simulated SDSS magnitudes of stars (small dots) and AGN (heavy dots) in 10 deg² at the North Galactic Pole (Fan 1999) with realistic log(F _x/F_opt) for these objects from the EMSS (Stocke et al. 1991). The known SDSS color degeneracy for stars and Quasi-Stellar Objects (QSOs) in the range 2.5<z<3 is completely removed by combining 2 SDSS colors with X-ray information.

$r^{\prime} - i^{\prime}$

LINK TO POSTSCRIPT FILE for Figure 16