The Spitzer c2d Survey of Large, Nearby, Interstellar Clouds: VI. Perseus Observed with MIPS L. M. Rebull1, K. R. Stapelfeldt2, N. J. Evans II7, J. K. Jørgensen8, P. M. Harvey7, T. Y. Brooke6, T. L. Bourke8, D. L. Padgett1, N. L. Chapman3, S.-P. Lai3,4,5, W. J. Spiesman7, A. Noriega-Crespo1, B. Mer´ın12, T. Huard8, L. E. Allen8, G. A. Blake9, T. Jarrett10, D. W. 7 Koerner11, L. G. Mundy3, P. C. Myers8, A. I. Sargent6, E. F. van Dishoeck12, Z. Wahhaj11, 0 K. E. Young7,13 0 2 n a ABSTRACT J 4 2 We present observations of 10.6 square degrees of the Perseus molecular cloud 1 at 24, 70, and 160 µm with the Spitzer Space Telescope Multiband Imaging v Photometer for Spitzer (MIPS). The image mosaics show prominent, complex 1 1 extended emission dominated by illuminating B stars on the East side of the 7 1 cloud, and by cold filaments of 160 µm emission on the West side. 0 7 0 / 1Spitzer Science Center/Caltech, M/S 220-6, 1200 E. California Blvd., Pasadena, CA 91125 h ([email protected]) p o- 2Jet Propulsion Laboratory,MS 183-900,California Institute of Technology, Pasadena,CA 91109 r t 3Department of Astronomy, University of Maryland, College Park, MD 20742 s a : 4Institute of Astronomy and Department of Physics, National Tsing Hua University, Hsinchu 30043, v Taiwan i X 5 Academia Sinica Institute of Astronomy and Astrophysics, P.O. Box 23-141,Taipei 106, Taiwan r a 6Division of Physics, Mathematics, and Astronomy, MS 105-24, California Institute of Technology, Pasadena,CA 91125 7DepartmentofAstronomy,UniversityofTexasatAustin,1UniversityStationC1400,Austin,TX78712 8Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, MS42, Cambridge, MA 02138 9DivisionofGeologicalandPlanetarySciences,MS150-21,CaliforniaInstitute ofTechnology,Pasadena, CA 91125 10 Infrared Processing and Analysis Center, California Institute of Technology, Pasadena,CA 91125 11Department of Physics and Astronomy, Northern Arizona University, NAU Box 6010, Flagstaff, AZ 86011-6010 12Leiden Observatory, PO Box 9513, NL 2300 RA Leiden, The Netherlands 13Department of Physical Sciences, Nicholls State University, Thibodaux, Louisiana 70301 – 2 – Of 3950 point sources identified at 24 µm, 1141 have 2MASS counterparts. A quarter of these populate regions of the K vs. K −[24] diagram that are distinct s s fromstellarphotospheresandbackgroundgalaxies, andthusarelikelytobecloud members with infrared excess. Nearly half (46%) of these 24 µm excess sources are distributed outside the IC 348 and NGC 1333 clusters. NGC 1333 shows the highest fraction of stars with flat or rising spectral energy distributions (28%), whileClassIISEDsaremostcommoninIC348. Theseresultsareconsistent with previous relative age determinations for the two clusters. A significant number of IRAS PSC objects are not recovered by Spitzer/MIPS, most often because the IRAS objects were confused by bright nebulosity. There is no evidence for 24 µm source variability to 10% between the ∼3-6 hours of our two observation epochs. The intercluster region contains several tightly clumped (r ∼0.1 pc) young stellar aggregates whose members exhibit a wide variety of infrared spectral en- ergy distributions characteristic of different circumstellar environments. One possible explanation is a significant age spread among the aggregate members, such that some have had time to evolve more than others. Alternatively, if the aggregate members all formed at roughly the same time , then remarkably rapid circumstellar evolution would be required to account for the association of Class I and Class III sources at ages . 1 Myr. We highlight important results for the HH 211 flow, where the bowshocks are detected at both 24 and 70 µm; and for the debris disk candidate BD +31◦643, where the MIPS data shows the linear nebulosity to be an unrelated interstellar feature. Our data, mosaics, and catalogs are available at the Spitzer Science Archive for use by interested members of the community. Subject headings: stars: formation – stars: circumstellar matter – stars: pre- main sequence – ISM: clouds – ISM: individual (IC 348, NGC 1333) – ISM: jets and outflows – infrared: stars – infrared: ISM 1. Introduction The Spitzer Space Telescope Legacy program“FromMolecular Cores to Planet-Forming Disks” (c2d; Evans et al. 2003) selected five large star-forming clouds for mapping with the Infrared Array Camera (IRAC, 3.6, 4.5, 5.8, and 8 µm; Fazio et al. 2004) and the Multiband Imaging Photometer for Spitzer (MIPS, 24, 70, and 160µm; Rieke et al. 2004). These clouds were selected to be within 350 pc, to have a substantial mass of molecular gas, and to have a range of cloud properties, thereby allowing studies of star formation in isolation, in groups, – 3 – and in clusters. The goals of this aspect of the c2d project (see Evans et al. 2003 for more information) include determining the stellar content of the clouds, the distributions of the youngest stars and substellar objects, and the properties of their disks and envelopes. All of these Spitzer cloud studies represent the first unbiased mid-infrared surveys across entire clouds at this sensitivity and spatial resolution, where in the past only targeted, small-field- of-view observations have been possible. Perseus is one of five nearby star forming clouds mapped with IRAC and MIPS by c2d, also including Chamaeleon II, Lupus I, III & IV, Ophiuchus and Serpens (see Evans et al. 2003 for an overview). Previous papers in this series presented IRAC observations of Serpens (Harvey et al. 2006), Perseus (Jørgensen et al. 2006, hereafter J06) and Chamaeleon II (Porras et al. 2007) and MIPS observations of Chamaeleon II (Young et al. 2005) and Lupus I, III and IV (Chapman et al. 2007). ThePerseusmolecularcloudisasourceofactivestarformationwithfewhigh-massstars, none earlier than early B. While nowhere near as chaotic as the Orion star forming region, it is also not as quiescent as the Taurus molecular cloud, and so provides an “intermediate” case study. IC 348 and NGC 1333 are the two densest and most famous star-forming clusters in this region, containing numerous stars <1-2 Myr old. Ongoing star formation is certainly occurring throughout the cloud, including in named regions such as L1448 and B5; there are deeply embedded Class 0 objects found outside the two major clusters (see, e.g., J06). The Perseus cloud is large enough that different parts of it may be at significantly different distances; see the discussion in Enoch et al. (2006). Following that paper, and J06, we take the distance to Perseus to be 250 pc, though we acknowledge that there may indeed be a substantial distance gradient across the cloud, or there might even be multiple pieces located at several different distances. The Perseus cloud is rich in both point sources and complex extended emission. IRAS observations revealed more than 200 apparent point sources and complex ISM structures, including a large ring thought to be an H II region, G159.6−18.5, excited by HD 278942 (Andersson et al. 2000). Ridge et al. (2006) argue that the ring is behind the main Perseus cloud. Hatchelletal.(2005)findviaasurvey at850and450µmahighdegreeofpoint-source clustering and filamentary structures throughout the cloud. Enoch et al. (2006) also find strong point-source clustering in their 1.1 mm continuum survey. The Spitzer/IRAC study of Perseus by J06 concluded (a) that there are significant numbers of stars being formed outside of the two main clusters (IC 348 and NGC 1333); (b) the fraction of Class I, Class II, and “flat spectrum” young stellar objects (YSOs) differs between the two rich clusters and the extended cloud population; and (c) that deeply embedded Class 0 objects are detected, with very red [3.6]−[4.5] colors (but not similarly red [5.8]−[8] colors). – 4 – MIPS observations at 24 µm (∼ 6′′ resolution), 70 µm (∼ 20′′ resolution), and 160 µm (∼ 40′′ resolution) can elucidate many aspects of the ongoing star formation in Perseus. Although the emission from stellar photospheres is falling rapidly at 24 µm, emission from circumstellar material makes many of the young cluster members still quite bright at 24 µm, so MIPS finds the young stars easily. Emission from the cloud itself becomes increasingly prominent in the MIPS 24, 70, and 160 µm bands, allowing dusty molecular material in the temperature range 120-20 K to be probed. MIPS reveals complex extended emission throughout the Perseus region at all three of its wavelengths. This paper presents MIPS data covering more than 10.5 square degrees in Perseus. We also use information obtained from the IRAC data for Perseus from J06. As a result of observational constraints (see §2 below) the IRAC data cover only about one-third the area of the MIPS data (IRAC covers 3.86 square degrees), so we use the IRAC data where possible, but large areas of our map have no IRAC data at all. The goalsof this paper are to present the MIPS data in a formatsimilar to that found in the other papers in the series, to discuss some of the high-level conclusions drawn from these data, and to highlight some of the interesting objects we have found. Because this paper is part of a series, there is synergy with both the papers that have gone before and those to come. Because theIRACdata, wheretheyexist, areusually importantforunderstanding the objects seen in the Perseus map, there are extensive references to J06; for example, SEDs for some objects discussed there were deferred to this paper for presentation. Similarly, there are references to future work throughout this paper. One such future paper will present a complete list of YSO candidates associated with Perseus using the combined IRAC and MIPS data, which is beyond the scope of this paper. As part of the c2d ancillary data, there has been a paper on the Bolocam 1.1 mm continuum survey of Perseus (Enoch et al. 2006) and one on the JCMT/SCUBA sub-mm maps from the COMPLETE team as compared to the Spitzer data (Jørgensen et al. 2006b). This paper can be broadly divided into three major parts. First we give the details of the observations, reductions, and source extraction (§2). This is followed by a presentation of the ensemble MIPS results for the entire Perseus cloud (§3). Finally, §4 gives a focused discussion of noteworthy stellar aggregates and individual young stars. The main results of our study are summarized in §5. – 5 – Fig. 1.— Location of MIPS coverage (small points are MIPS-24 detections), with the region of IRAC coverage (dashed line) indicated. The smaller squares (solid lines) indicate the regions defined to be IC 348 (left) and NGC 1333 (right). The definition of IC 348 includes the new objects recently found by Lada et al. (2006). Table 1. Summary of observations (program 178) field map center first epoch AORKEY second epoch AORKEY per1 3h47m05.0s,+32d38m23.0s 5780992 5787648 per2 3h44m30.7s,+32d06m08.1s 5781248 5787904 per3 3h42m34.5s,+31d55m33.0s 5781504 5788160 per4 3h40m39.0s,+31d37m53.0s 5781760 5788416 per5 3h37m42.0s,+31d16m41.0s 5782016 5788672 per6 3h33m34.0s,+31d08m57.0s 5782272 5788928 per7 3h31m10.6s,+30d51m12.0s 5782528 5789184 per8 3h29m10.0s,+31d11m00.0s 5782784 5789440 per9 3h30m54.0s,+30d00m48.0s 5783040 5789696 per10 3h28m26.0s,+30d39m55.0s 5783296 5789952 per11 3h26m11.0s,+30d32m03.0s 5798656 5790208 – 6 – 2. Observations, Data Reduction, and Source Extraction 2.1. Observations The MIPS observations of Perseus were conducted on 18-20 Sep 2004 and covered ∼10.5 square degrees; they were designed to cover the A =2 contour, which then by extension V completely covered the c2d IRAC map of Perseus. The center of this large map is roughly at α,δ (J2000) = 3h37m, 31◦11′30, or galactic coordinates l,b=160◦, −19.5◦, or ecliptic coordinates (J2000) 59.5◦, +11.5◦. These observations were part of Spitzer program id 178; “AORKEYs” labeling the datasets in the Spitzer Archive are given in Table 1. Fast scan maps were obtained at two separateepochs. Ateachepoch,thespacingbetweenadjacentscanlegswas240′′. Thesecond epochobservationwasoffset 125′′ inthecross-scandirectionfromthefirst, tofillinthe70µm sky coverage which was incomplete at each individual epoch. Furthermore, the second epoch scan was also offset 80′′ from the first along the scan direction, to maximize 160 µm map coverage in the combined epoch mosaic. These mapping parameters resulted in every part of the map being imaged at two epochs at 24 µm and only one epoch at 70 and 160 µm, with total integration times of 30 sec, 15 sec, and 3 sec at each point in the map (respectively). The 160 µm maps have some coverage gaps, and suffer from saturation (particularly in NGC 1333 and IC 348). Figure 1 shows the region of 3-band coverage with MIPS, and the 4-band coverage with IRAC. At about 10.5 square degrees, the MIPS observations cover a much larger area than the IRAC observations, which cover only ∼4 square degrees. The three main reasons for this apparent mismatch are entirely instrumental: (1) we are constrained by the ecliptic latitude to observe with scan legs in a particular orientation (the maximum rotation is about 10◦), (2) we are limited in the available choices for scan leg lengths, and (3) MIPS covers large areas very efficiently, so we can easily cover large areas with MIPS in much less time than for IRAC. The c2d MIPS observations were designed for even coverage (all to the same exposure time depth), independent of the GTO observations of IC 348 (Lada et al. 2006) and NGC 1333 (R. Gutermuth et al. in prep.). The GTO observations of these regions are not included in the discussion here, primarily to enable discussion of a catalog obtained to uniform survey depth. This is different from the c2d IRAC observations of Perseus discussed in J06, where the GTO observations provided the images for one of the two epochs. The two observation epochs were separated by 3 to 6 hrs to permit asteroid removal in this relatively low ecliptic latitude (+11–12◦) field. Indeed, by comparing the 24 µm maps obtained at the two epochs, at least 100 asteroids are easily visible, ranging in flux density from at least as faint as 0.6 to as bright as 30 mJy. Some asteroids are clearly visible even – 7 – at 70 µm. The asteroids in this and other c2d cloud maps will be discussed further in K. Stapelfeldt et al., in preparation. WestartedwiththeSSC-pipeline-produced basiccalibrateddata(BCDs), versionS11.4. For a description of the pipeline, see Gordon et al. (2005). As in Chapman et al. (2006), each MIPSchannelwasthenprocesseddifferentlyandthereforeisdiscussedseparatelybelow. Mo- saics and source catalogs from these data were delivered back to the SSC for distribution; see http://ssc.spitzer.caltech.edu/legacy/ for additional information. Multiple deliveries were made; the data discussed here were part of the 2005 data delivery. Figures 2, 3, and 4 show the individual mosaics by channel, and 5 shows a 3-color image with all three channels included. (For anindication of where “famous” regions are, please see Fig.16,whichindicatesseveralobjectshighlightedfordiscussion below.) Thereissubstantial extended emission in all three MIPS channels throughout the MIPS maps. In the 70 and 160 µm channels, the MIPS instrument uses internal stimulator flashes to calibrate the data (for more information, see the Spitzer Observer’s Manual, available at the SSC website1). For most of a scan leg, the correct calibration can be obtained via an interpolation. On the ends of scan legs, it necessarily must use extrapolation solutions. When the ends of the scan legs run across particularly bright emission, as they do in some cases here (particularly in the “ring” of bright emission), the absolute calibration is not as good as it is in darker regions. Fluxes obtained in these regions have larger errors than the rest of the map. The IC 348 and NGC 1333 regions are encompassed by the overall Perseus cloud map dataset. To explore difference between the memership of these two clusters and the more broadly distributed Perseus young stellar population, we have chosen to consider the IC 348 and NGC 1333 stellar populations separately from the rest of the cloud; see Figure 1. We have defined the regions belonging to the clusters based on the surface density of 24 µm sources. The region we define to be IC 348 is given by a box bounded by the coordinates α=55.8◦ to 56.5◦ (3.720h to 3.767h, or 03h43m12.0s to 03h46m00.0s), δ=31.8◦ to 32.4◦ (or +31◦48′00.0′′ to+32◦24′00.0′′). This regionislarger thanwhat hashistorically been assumed to encompass the cluster, and large enough (0.36 square degrees) to include the ∼300 likely cluster members (with the new likely members) found by Lada et al. (2006), but not so large that it includes substantial numbers of likely field members. (Note that this box includes most of the new members found by Cambresy et al. 2006, but does not include their farthest southwest part of the cluster.) The region we define to be NGC 1333 is given by a box bounded by the coordinates α=52◦ to 52.5◦ (3.467h to 3.500h, or 03h28m00.0s to 03h30m00.0s), δ=31.1◦ to 31.5◦ (or +31◦06′00.0′′ to +31◦30′00′′); this region is 0.17 square 1http://ssc.spitzer.caltech.edu/ – 8 – degrees. In both cases, for ease of comparison, these regions are the same as those used in J06. 2.2. MIPS-24 As discussed in Chapman et al. (2006), standard c2d pipeline processing on the S11 BCDs (2005 delivery) was used for MIPS-24 (Evans et al. 2005; see also Young et al. 2005). In summary, the c2d reduction starts with the pipeline-produced BCDs, and then further processes them to remove artifacts, e.g., “jailbars” near bright sources. A mosaic was then constructed from the entire data set using the SSC MOPEX software (Makovoz & Marleau 2005); see Figure 2. Sources were extracted from the mosaic and bandmerged into the catalog along with 2MASS (Skrutskie et al. 2006) JHK and IRAC-1,2,3,4 measurements. s For more details on this process, please see Chapman et al. (2006). The uncertainty on the flux densities derived at 24 µm is estimated to be 15%. The high-quality catalog we assembled consisted of all detections at MIPS-24 from the last 2005 delivery where the c2d catalog detection quality flag (see the 2005 c2d Delivery Document, available online linked from http://ssc.spitzer.caltech.edu/legacy/) was ‘A’ or ‘B,’ which translates to a signal-to-noise ratio of >5, and where the object was detected at both epochs. While resulting in a shallower survey than would be possible using other combinations of flags, this ensures that no asteroids are included in the catalog. The extrac- tion pipeline flags some objects as extended (“imtype” flag); no filter was imposed on these extended objects to create the catalog we used, but only 8 of the objects in our catalog are flagged as extended. There were 3950 total point sources detected at 24 µm meeting our criteria, ranging from 0.603 to 3530 mJy (comparable to the saturation limit; see below). The source surface density is about 370 sources per square degree. The zero point used to convert these flux densities to magnitudes was 7.14 Jy, based on the extrapolation from the Vega spectrum as published in the MIPS Data Handbook. About 30% of these objects had identifiable 2MASS counterparts at K . s The faint limit of the catalog of 24 µm sources is a function of the nebular brightness across the field, but what might be less obvious is that the saturation limit for point sources with MIPS-24 is also a function of location in the cloud because the total flux density registered by the detector is that dueto the point source itself plus any surrounding extended nebular emission. Because the extended emission at 24 µm varies from ∼450 MJy sr−1 in IC 348 to ∼100 MJy sr−1 in NGC 1333 to .1 MJy sr−1 in the darker parts of the cloud, the – 9 – Fig. 2.— Mosaic of Perseus map at 24 µm. The reverse greyscale colors correspond to a logarithmic stretch of surface brightnesses. – 10 – completeness of the 24 µm catalog at both the bright and faint limits is a function of location in the cloud. For example, as will be seen in the source counts discussion (§3.1), there are fewer faint sources in the clusters than in the field, and fewer bright sources in NGC 1333 than in IC 348. This may indeed be entirely due to the brightness of the background. An additional issue when considering completeness is the resolution; the resolution of MIPS-24 (∼ 6′′, 2.55′′ pixel size) is poorer than IRAC or 2MASS (∼ 2′′). Source multiplicity and confusion may also affect the completeness of the catalog, particularly in dense regions such as the clusters. 2.3. MIPS-70 To reduce the MIPS-70 data, we started with the automated pipeline-produced BCDs. The SSC produces two sets of BCDs; one is simply calibrated, and the other has spatial and temporal filters applied that attempt to remove instrumental signatures in an automated fashion. These filtered BCDs do not conserve flux for extended emission, nor for bright point sources, but they do conserve flux for the fainter point sources. We started with both the filtered and unfiltered S11 BCD products. Then, we mosaicked these individual BCDs into one filtered and one unfiltered mosaic using the MOPEX software. We resampled the pixels to be 4′′ square (smaller than the native pixel scale of ∼ 10′′) to better enable source extraction. The unfiltered mosaic is presented in Figure 3 to better show the extended emission. We defined the point response function (PRF) from clean and bright point sources selected from this large mosaic, and then using this, performed point source detection and extractionusing theAPEX-1 frameoptionofMOPEX. Theinitially-producedsourcelist was cleaned for instrumental artifacts via manual inspection of the 70 µm image and comparison to the 24 µm image; e.g., if there was some question as to whether a faint object seen at 70 µm was real or an instrumental artifact, and comparison to the 24 µm image revealed a 24 µm source, then the 70 µm object was retained as a real source. This catalog also has the same limitations as was found at 24 µm; the brightness of the nebulosity drowns out the faintest objects in the clusters, and contributes to saturation of the brightest point-source objects (particularly in the clusters) as well. And, the resolution at 70 µm (∼ 20′′) is coarser than it is at 24 µm (∼ 6′′), which complicates source matching to K and source extraction s in confused regions such as the clusters. For all of these reasons, the 70 µm catalog is not necessarily complete and unbiased, particularly in the regions of bright ISM and/or the faintest end. Based on a comparison of PRF and aperture photometry fluxes, we empirically de-