DOG Luminosities and Dust Properties

Infrared Luminosities and Dust Properties of Dust-Obscured Galaxies

R. S. Bussmann11affiliation: Steward Observatory, Department of Astronomy, University of Arizona, 933 N. Cherry Ave., Tucson, AZ 85721; , Arjun Dey22affiliation: National Optical Astronomy Observatory, 950 N. Cherry Ave., Tucson, AZ 85719 , C. Borys33affiliation: Herschel Science Center, California Institute of Technology, Pasadena, CA 91125 , V. Desai44affiliation: Spitzer Science Center, California Institute of Technology, MS 220-6, Pasadena, CA 91125 , B. T. Jannuzi22affiliation: National Optical Astronomy Observatory, 950 N. Cherry Ave., Tucson, AZ 85719 , E. Le Floc’h55affiliation: Spitzer Fellow, Institute for Astronomy, University of Hawaii, Honolulu, HI 96822 , J. Melbourne66affiliation: Division of Physics, Math and Astronomy, California Institute of Technology, Pasadena, CA 91125 , K. Sheth44affiliation: Spitzer Science Center, California Institute of Technology, MS 220-6, Pasadena, CA 91125 , B. T. Soifer44affiliation: Spitzer Science Center, California Institute of Technology, MS 220-6, Pasadena, CA 91125 66affiliation: Division of Physics, Math and Astronomy, California Institute of Technology, Pasadena, CA 91125

We present SHARC-II 350m imaging of twelve 24m-bright (mJy) Dust-Obscured Galaxies (DOGs) and CARMA 1mm imaging of a subset of 2 DOGs. These objects are selected from the Boötes field of the NOAO Deep Wide-Field Survey. Detections of 4 DOGs at 350m imply infrared (IR) luminosities which are consistent to within a factor of 2 of expectations based on a warm dust spectral energy distribution (SED) scaled to the observed 24m flux density. The 350m upper limits for the 8 non-detected DOGs are consistent with both Mrk 231 and M82 (warm dust SEDs), but exclude cold dust (Arp 220) SEDs. The two DOGs targeted at 1mm were not detected in our CARMA observations, placing strong constraints on the dust temperature:  K. Assuming these dust properties apply to the entire sample, we find dust masses of . In comparison to other dusty galaxy populations such as sub-millimeter galaxies (SMGs) and other Spitzer-selected high-redshift sources, this sample of DOGs has higher IR luminosities ( vs. for the other galaxy populations) that are driven by warmer dust temperatures (35-60 K vs. 30 K) and lower inferred dust masses ( vs. ). Wide-field Herschel and SCUBA-2 surveys should be able to detect hundreds of these power-law dominated DOGs. We use existing Hubble Space Telescope and Spitzer/IRAC data to estimate stellar masses of these sources and find that the stellar to gas mass ratio may be higher in our 24m-bright sample of DOGs than in SMGs and other Spitzer-selected sources. Although much larger sample sizes are needed to provide a definitive conclusion, the data are consistent with an evolutionary trend in which the formation of massive galaxies at involves a sub-millimeter bright, cold-dust and star-formation dominated phase followed by a 24m-bright, warm-dust and AGN-dominated phase.

Subject headings:
galaxies: evolution — galaxies: fundamental parameters — galaxies: high-redshift — submillimeter

1. Introduction

In the local universe, the most bolometrically luminous galaxies are dominated by thermal emission from dust which absorbs ultra-violet (UV) and optical light and re-radiates it in the infrared (IR) (Soifer et al. 1986). While rare locally, these ultra-luminous IR galaxies (ULIRGs) are more common at high redshift (e.g., Franceschini et al. 2001; Le Floc’h et al. 2005; Pérez-González et al. 2005). Studies combining the improved sensitivity in the IR of the Spitzer Space Telescope with wide-field ground-based optical imaging have identified a subset of this ULIRG population that is IR-bright but also optically faint (Yan et al. 2004; Houck et al. 2005; Weedman et al. 2006b; Fiore et al. 2008; Dey et al. 2008). In particular, Dey et al. (2008) and Fiore et al. (2008) present a simple and economical method for selecting these systems based on -band and 24m Multiband Imaging Photometer for Spitzer (MIPS; Rieke et al. 2004) data. Dey et al. (2008) employ a color cut of (Vega magnitudes; ) to identify objects they call Dust-Obscured Galaxies (DOGs). Applied to the 8.6 deg Boötes field of the NOAO Deep Wide-Field Survey (NDWFS) that has uniform MIPS 24m coverage for mJy , this selection yields a sample of 2600 DOGs, or 302 deg.

The extreme red colors and number density of the DOGs imply that they are undergoing a very luminous, short-lived phase of activity characterized by vigorous stellar bulge and nuclear black hole growth. Spectroscopic redshifts determined for a sub-sample of DOGs using the Deep Imaging Multi-Object Spectrograph (DEIMOS; Faber et al. 2003) and the Low Resolution Imaging Spectrometer (LRIS; Oke et al. 1995) on the telescopes of the W. M. Keck Observatory (59 DOGs), as well as the Infrared Spectrometer (IRS; Houck et al. 2004) on Spitzer (47 DOGs) have revealed a redshift distribution centered on with a dispersion of (Dey et al. 2008).

While DOGs are rare, they are sufficiently luminous (% of DOGs with spectroscopic redshifts have ) that they may contribute up to one-quarter of the total IR luminosity density from all galaxies (and over half that from all ULIRGs at ) and may be the progenitors of the most luminous (4) present-day galaxies (Dey et al. 2008; Brodwin et al. 2008). Thus far, the efforts to estimate the IR luminosities of DOGs have primarily relied upon spectroscopic redshifts and the observed 24m flux density. Dey et al. (2008) use an empirical relation between the rest-frame 8m luminosity (computed from the observed 24m flux density) and the IR luminosity, derived by Caputi et al. (2007). However, there is evidence from sources with mJy that methods based on only the 24m flux density can overestimate the IR luminosity by factors of 2-10 (Papovich et al. 2007). Results from deep 70m and 160m imaging of a sub-sample of 24m-bright DOGs are consistent with this, favoring hot-dust dominated spectral energy distribution (SED) templates like that of Mrk 231 (Tyler et al. 2009) which lead to estimates of the IR luminosity that are on the low end of the range in conversion factors adopted in Dey et al. (2008).

In this paper, we present 350m and 1mm photometry of a sample of DOGs whose mid-IR spectral features (silicate absorption, power-law SEDs) suggest the presence of a strong active galactic nucleus (AGN). The primary goals of this study are to measure their IR luminosities and constrain their dust properties, in particular the dust masses and temperatures. We also estimate stellar masses for the sources in the sample using published Hubble Space Telescope (HST) data and Spitzer InfraRed Array Camera (IRAC) catalogs from the Spitzer Deep Wide-Field Survey (SDWFS; see Ashby et al. 2009). Comparison of the stellar and dust masses potentially allows us to place constraints on the evolutionary status of these sources.

In section 2, we present the details of the observations. Section 3 presents the DOG SEDs from 0.4m to 1mm and IR luminosity measurements, constraints on the dust temperature, and dust and stellar mass estimates. In section 4, we compare our results with similar studies of sub-millimeter galaxies (SMGs) and Spitzer-selected sources from the eXtragalactic First Look Survey (XFLS) and Spitzer Wide InfraRed Extragalactic (SWIRE) survey. We present our conclusions in section 5.

Throughout this paper we assume a cosmology where 70 km s Mpc, , and .

2. Observations

2.1. Sample Selection

Figure 1.— color vs. 24m magnitude for DOGs in the NDWFS Boötes field. Bottom and top abscissae show the 24m magnitude and flux density, respectively. Left and right ordinates show the color in magnitudes and the flux density ratio, respectively. Black dots and upward arrows show the full sample of DOGs, with and without an -band detection, respectively. The subsample studied in this paper is represented by red circles (open symbols show sources undetected in the -band data). Two sources observed by CARMA at 1mm are highlighted by a black square. This plot demonstrates that the sample studied in this paper probes the 24m-bright DOGs over a wide range of colors.

Dey et al. (2008) identified 2603 DOGs in the 8.6 deg NDWFS Boötes field, selecting all 24m sources satisfying (Vega mag) and mJy. We identified 12 DOGs with spectroscopic redshifts for follow-up 350m imaging (see Figure 1) with the second-generation Submillimeter High Angular Resolution Camera (SHARC-II) at the Caltech Sub-millimeter Observatory (CSO). These targets were selected to have bright 24m flux densities ( mJy) and a power-law dominated mid-IR SED (based on Spitzer/IRAC and 24m MIPS photometry; for details, see section 3.1.2 in Dey et al. 2008). Using the deeper IRAC observations from SDWFS (Ashby et al. 2009), the fraction of DOGs qualifying as power-law sources ranges from at  mJy to at  mJy. As shown in Figure 1, our sample spans a broad range in color (). Spitzer/IRS spectroscopic redshifts have been obtained for these sources based on the 9.7m silicate absorption feature. Power-law continua and silicate absorption are typical features of AGN-dominated systems (Donley et al. 2007; Weedman et al. 2006a; Polletta et al. 2008; Brand et al. 2008). We note that such systems often exhibit intense star-formation concurrent with the growth of a super-massive central black hole (e.g., Wang et al. 2008).

Details of our observations are presented in Table 1. The effective integration time (column 8 of table 1) represents the time necessary to reach the same noise level given a completely transparent atmosphere (see Coppin et al. 2008, for details).

We observed two of the twelve DOGs at 1mm using the Combined Array for Research in Millimeter-wave Astronomy (CARMA) interferometer to search for thermal emission from cold dust particles. These were primarily selected to have robust 350m detections to enhance the probability of detection at 1mm and, in the event of a non-detection, allow useful constraints to be placed on the dust properties. The two targets observed with CARMA are SST24 J142827.2+354127 (S2) and SST24 J143001.9+334538 (S3). Table 2 presents the date and integration time of the CARMA observations.

RA DEC aaActual on-source integration time. bbEffective integration time for a transparent atmosphere (Coppin et al. 2008).
Source Name ID (J2000) (J2000) UT Year-Month (hr) (min) (mJy)
SST24 J142648.9+332927 S1 14:26:48.970 +33:29:27.56 2.00ccRedshift from Spitzer/IRS (Higdon et al. in prep). 2006-Apr 1.1 4.0 20
SST24 J142827.2+354127 S2 14:28:27.190 +35:41:27.71 1.293ddRedshift from Keck DEIMOS (Desai et al. 2006). 2005-Apr/2006-Apr/2007-May 3.3 11.9 10
SST24 J143001.9+334538 S3 14:30:01.910 +33:45:38.54 2.46eeRedshift from Spitzer/IRS (Houck et al. 2005). 2005-Apr/2006-Apr/2006-May 3.6 11.4 8
SST24 J143025.7+342957 S4 14:30:25.764 +34:29:57.29 2.545ffRedshift from Keck DEIMOS (Dey et al., in prep.). 2006-Apr 1.0 2.0 26
SST24 J143135.2+325456 S5 14:31:35.309 +32:54:56.84 1.48ccRedshift from Spitzer/IRS (Higdon et al. in prep). 2007-May 1.2 2.0 30
SST24 J143325.8+333736 S6 14:33:25.884 +33:37:36.90 1.90ccRedshift from Spitzer/IRS (Higdon et al. in prep). 2006-Apr 0.3 1.2 28
SST24 J143411.0+331733 S7 14:34:10.980 +33:17:32.70 2.656ggRedshift from Keck LRIS (Dey et al. 2005). 2005-Jan 3.8 12.7 8
SST24 J143447.7+330230 S8 14:34:47.762 +33:02:30.46 1.78eeRedshift from Spitzer/IRS (Houck et al. 2005). 2007-May 1.8 10.5 10
SST24 J143508.4+334739 S9 14:35:08.518 +33:47:39.44 2.10ccRedshift from Spitzer/IRS (Higdon et al. in prep). 2006-Apr 1.0 0.5 35
SST24 J143539.3+334159 S10 14:35:39.364 +33:41:59.13 2.62eeRedshift from Spitzer/IRS (Houck et al. 2005). 2005-Apr/2006-Apr 2.5 3.1 15
SST24 J143545.1+342831 S11 14:35:45.137 +34:28:31.42 2.50ccRedshift from Spitzer/IRS (Higdon et al. in prep). 2006-Apr 0.8 3.2 22
SST24 J143644.2+350627 S12 14:36:44.269 +35:06:27.12 1.95eeRedshift from Spitzer/IRS (Houck et al. 2005). 2006-Apr/2006-May 2.1 6.7 10
Table 1SHARC-II 350m Observations
ID UT Year-Month (hr)aaOn-source integration time.
S2 2008-April/May 10.3
S3 2008-April/May 7.5
Table 2CARMA 1mm Observations

2.2. Sharc-Ii 350m Imaging and Photometry

The SHARC-II observations of the 12 target DOGs were carried out over the course of five separate observing runs from 2005 January to 2007 May. Data were collected only when the atmospheric opacity was low and conditions were stable (). Pointing, focus checks, and calibration were performed every hour, using ULIRG Arp 220 as a calibrator (Jy). Other secondary calibrators (CIT6, CRL2688, 3C345) were used occasionally to verify the Arp 220 calibrations. A non-connecting Lissajous pattern was used to modulate the telescope pointing with amplitudes of 15-20 and periods of 10-20 s. The observations made use of the CSO Dish Surface Optimization System to optimize the dish surface accuracy and beam efficiency (Leong et al. 2006).

Data were reduced using the Comprehensive Reduction Utility for SHARC-II (CRUSH) software package with the ‘deep’ option to optimize the signal-to-noise ratio (S/N) for faint ( 100 mJy) point sources (Kovács 2006). The output map has a pixel scale of 162 pixel, and is smoothed with a 9 gaussian beam, resulting in an effective image FWHM of 124.

A 20 diameter aperture was used for photometry to compute the instrumental flux density of each source. The sky level and photometric uncertainty were computed by measuring the mean and RMS in 10 off-source 20 diameter apertures. The same procedure was applied to the calibration images, and a scaling factor was derived that converts the instrumental flux density to a physical flux density (using this method, no subsequent aperture correction is required as long as both the science and calibration targets are unresolved and measured in the same aperture).

The aperture photometry is consistent with peak flux density measurements in all but one source, SST24 J142648.9+332927. This source has a radial profile that is significantly more extended than the point spread function of the final map, which results in the peak flux underestimating the aperture flux density measurement. The extended structure in the image is more likely to be noise than signal, so in this case we report the peak flux measurement, which is formally a non-detection.

Flux boosting of low S/N sources can be an important effect in wide-field surveys where source positions are not known a priori (e.g., Coppin et al. 2005). However, because we know our source positions at the level (from MIPS and IRAC centroids), flux boosting is not a significant effect, and so we do not apply any such corrections to our measurements. Our approach follows that adopted by Laurent et al. (2006) and Kovács et al. (2006) in their 350m follow-up imaging of SMGs.

2.3. CARMA 1mm Imaging and Photometry

The CARMA observations were obtained between 2008 April 7 and May 1 in the C-array configuration (beamsize is sq. arcsec). A total of 7.5 hours and 10.3 hours of integration time in good 1mm weather conditions were spent on sources SST24 J143001.9+334538 (S3) and SST24 J142827.2+354127 (S2), respectively. These sources were selected primarily because of their robust (S/N) detections at 350m. In addition, source S3 is detected with Spitzer/MIPS at 70m and 160m (Tyler et al. 2009), while S2 is the subject of a detailed spectroscopic study (Desai et al. 2006).

System temperatures were in the range 250-400 K. A correlator configuration was used with three adjacent 1531 MHz bands centered on 220 GHz, the frequency at which the CARMA 1mm receivers are most sensitive. The quasar J1310+323 (chosen for its spatial proximity) was observed every 15 minutes for amplitude and phase calibration. Quasars 3C 273 and MWC 349 were used for pointing, pass-band calibration, and flux calibration.

Data were reduced using the Multichannel Image Reconstruction, Image Analysis, and Display (MIRIAD) software package (Sault et al. 1995). Visual inspection of visibilities as a function of baseline length allowed us to identify and flag spurious data. A cleaned map was generated for each track of integration time (ranging from 1 to 5 hours in length) and these tracks were coadded together using a weighted mean to obtain a final image of these sources. No detections were found in either case. Both sources are unresolved in the IRAC images, and NICMOS/F160W imaging of S3 indicates a size of , implying that it is very unlikely any emission was resolved out by the interferometer. The imstat routine from MIRIAD was used to determine the noise level in the co-added images where we expected to see the source. Table 3 shows the photometry from 24m to 1mm, where available. Non-detections are given as 3- upper limits.

2.4. Optical, near-IR, mid-IR, and far-IR Photometry

The optical and near-IR photometry used in this paper to estimate stellar masses are based on high spatial-resolution HST data (WFPC2/F606W, ACS/F814W, and NIC2/F160W) published in Bussmann et al. (2009). The HST data allow the separation of an unresolved nuclear component (flux on scales kpc) from a more spatially extended component. Because the AGN contribution in the rest-frame UV to optical is uncertain, the photometry of the extended component is used here (measured with 2 diameter apertures) to ensure that our stellar mass estimates are not biased by the presence of an obscured AGN (for details on how the extended component is computed, see Bussmann et al. 2009). Additionally, 4 diameter aperture photometry in the optical (, , and ) from the NDWFS is shown in the SEDs of the objects in this sample (details on how the photometry is computed may be found in Bussmann et al. 2009).

The mid-IR photometry used in this paper are from the publicly available Data Release 1.1 (DR1.1) catalogs from the SDWFS IRAC coverage of the Boötes field (Ashby et al. 2009). The SDWFS catalogs incorporate the earlier IRAC Shallow Survey of the Boötes field undertaken by the IRAC guaranteed time observation (GTO) programs (Eisenhardt et al. 2004). We identified IRAC counterparts of the DOGs in this paper from the SDWFS catalogs using a 3 search radius centered on the 24m position (the MIPS 24m 1- positional uncertainty is ). All of the DOGs in this paper have IRAC counterparts, detected at in all four IRAC channels. We use the 4 (rather than the 6) diameter aperture photometry from the DR1.1 SDWFS catalog to reduce contamination from nearby sources. We note that aperture corrections derived from isolated, bright stars have been applied to the SDWFS catalogs.

Finally, 24, 70, and 160m data over 8.61 deg of the Boötes field are available from GTO programs. The data were reduced by the MIPS GTO team and reach 1 rms depths of Jy, 5 mJy, and 18 mJy at 24, 70, and 160m, respectively. Details of the GTO surveys, such as mapping strategy, data reduction, and source catalogs, will be discussed elsewhere. In addition, several of the DOGs in this paper were targeted for deeper MIPS photometric observations by Spitzer General Observer program 20303 (P.I. E. LeFloc’h), and the results are reported in Tyler et al. (2009). We use the Tyler et al. (2009) measurements where they are available.

bbPhotometry from Westerbork Synthesis Radio Telescope imaging de Vries et al. (2002).
Source Name (Vega mag) (mJy) (mJy) (mJy) (mJy) (mJy) (mJy)
S1 16.1 2.330.07
S2 15.7 10.550.13 45ccPhotometry from Desai et al. (2006). 7413 1.5
S3 17.4 3.840.06 9.32.3ddPhotometry from Tyler et al. (2009). 6511ddPhotometry from Tyler et al. (2009). 4113 1.8
S4 15.1 2.470.05 81
S5 14.5 1.510.05 100
S6 15.4 1.870.06 137
S7 12.4ee-band photometry includes diffuse emission from Ly- nebula (Dey et al. 2005). 0.860.05 25ffPhotometry from Dey et al. (2005). 90ffPhotometry from Dey et al. (2005). 3713
S8 16.8 1.710.04 4512
S9 15.3 2.650.08 150
S10 16.7 2.670.06 8.1ddPhotometry from Tyler et al. (2009). 38ddPhotometry from Tyler et al. (2009). 50
S11 16.0 1.950.05 60
S12 15.4 2.340.05 9.12.5 4312 34
Table 3PhotometryaaUpper limits quoted are 3 values.
Figure 2.— SEDs of 5 DOGs detected by SHARC-II at 350m. Dotted, dashed, and dot-dashed lines show the Mrk 231, Arp 220, and M82 template SEDs, respectively, placed at the appropriate redshift and scaled to match the observed 24m flux density. Red horizontal lines show the 5- sensitivity limits (ignoring confusion) from the planned wide-field Herschel surveys at 250, 350, and 500m, and SCUBA-2 surveys at 850m. The cool dust SED of Arp 220 significantly overpredicts the 350m flux density in all cases. The warm dust SED of M82 provides a better fit in the far-IR, but Mrk 231 provides the best fit in both the far-IR and the optical.
Figure 3.— Same as Figure 2, except showing SEDs of 7 DOGs not detected by SHARC-II at 350m. The limits at 350m are all inconsistent with the Arp 220 SED. The 250m channel of Herschel should be very efficient for detecting power-law DOGs in wide-field surveys, assuming that Mrk 231 is an appropriate representation of the far-IR SED.

3. Results

In this section, we present SEDs from 0.4m to 1mm for each source in our sample and compare with local starburst (M82) and ULIRG (Arp 220 and Mrk 231) templates. Our approach is to artificially redshift the local galaxy templates and normalize them to match the DOG photometry at observed-frame 24m. This allows a simple, qualitative comparison of DOGs and galaxies with properties ranging from warm dust, star-formation dominated (M82), to cool dust, star-formation dominated (Arp 220), to warm dust, AGN-dominated (Mrk 231). We will use the SED that provides the best fit over the sampled wavelength range (observed optical through sub-mm) to estimate the IR luminosities of the DOGs.

Later in this section, we use our limits at 1mm from CARMA to place constraints on the dust temperatures and limits on the dust masses of DOGs. Finally, we use HST and Spitzer/IRAC data to estimate stellar masses of DOGs.

3.1. Qualitative SED Comparison

Figures 2 and 3 show the SEDs of the sources, divided respectively into those with and those without detections at 350m. Note that the rest-frame UV photometry for source S7 is contaminated by emission from nearby sources and will be treated in more detail in a future paper (Prescott et al., in prep.).

Overplotted in each panel are M82 (Silva et al. 1998)111We use a slightly updated SED obtained from, Mrk 231 (Chary 2008, private communication), and Arp 220 (Rieke et al. 2008) templates, placed at the appropriate redshift and scaled to match the flux density observed in the MIPS 24m band. The scaling factors derived for the three templates range over 200-900, and 2-10, 70-700, respectively (the deep silicate absorption feature in Arp 220 and the strong PAH emission feature of M82 make the scaling factors closer to each other than a simple estimate based on the ratio of the IR luminosities would imply). For the Arp 220 and Mrk 231 templates, we have interpolated the spectrum in the UV to match Galaxy Evolution Explorer photometry (in the case of Arp 220) and International Ultraviolet Explorer data as well as HST Faint Object Spectrograph data (in the case of Mrk 231; Hutchings & Neff 1987; Gallagher et al. 2002).

These templates were chosen because they sample a range of dust temperatures and AGN/starburst contributions. M82 is one of the closest (Mpc) galaxies undergoing a starburst, as it was triggered by a recent interaction with M81. Although it is less luminous than DOGs (; 222 is the luminosity integrated over 8-1000m Sanders et al. 2003), its nucleus is dominated by a warm dust component (K; Hughes et al. 1994). Arp 220 is a nearby (Mpc) ULIRG (; Sanders et al. 2003) dominated by cold dust (K; Rigopoulou et al. 1996). Mrk 231 is another nearby (Mpc) ULIRG (; Sanders et al. 2003), but has a warm dust (K; Yang & Phillips 2007) SED dominated by an obscured AGN.

Qualitatively, the Mrk 231 template provides a much better fit than the Arp 220 template to the 350m photometry in every case. M82 fits the 24m and 350m photometry reasonably well (although not as well as Mrk 231), but it fares poorly in the mid-IR and optical, where a strong stellar component in M82 is not seen in the DOGs in this sample (which are dominated by a power-law in the mid-IR). Additionally, M82 shows strong PAH emission which is not seen in the power-law DOGs.

The red horizontal bars in Figures 2 and 3 show 5- limits (ignoring confusion) from planned wide-field (deg) surveys with the Herschel Space Observatory (shown for the channels at 250, 350, and 500m) and with the Sub-mm Common-User Bolometer Array-2 (SCUBA-2) instrument at 850 m. Most of the power-law DOGs studied in this paper have SEDs that peak around observed-frame 250m, which is where the Herschel wide-field maps will be the deepest (5- limit of 14 mJy). If all of the 24m-bright DOGs are detected at 250m in the two wide-field surveys that are planned to reach the depths assumed here (Lockman Hole east, 11 deg; Extended Chandra Deep Field South, 8 deg), then a total of power-law DOGs should be detected in the 250m Herschel catalogs of these two fields. The SCUBA-2 surveys of these fields should be deep enough (5- limit of 3.5 mJy at 850m) to detect many of these sources, allowing dust temperature constraints to be placed on a statistically significant sample of these rare, important objects.

Figure 4 shows all of the DOG SEDs on the same plot, normalized by the rest-frame 8m flux density (which is estimated from the observed-frame 24m flux density by assuming a power law of the form , where ). The SEDs M82, Arp 220, Mrk 231, and a composite SMG template SED spanning mid-IR to sub-mm wavelengths are also shown. The composite SMG template is derived from bright (mJy) SMGs from the Great Observatories Origins Deep Survey North (GOODS-N) field with mid-IR spectra (Pope et al. 2008b).

Figure 4.— Optical through sub-mm SEDs of DOGs in the SHARC-II sample. Flux densities have been normalized by the rest-frame 8m flux density, computed from the observed 24m flux density. Of the three local galaxy templates shown, Mrk 231 provides the best fit over the rest-frame UV through sub-mm range because it has a warm dust SED (unlike Arp 220) and because it lacks a strong stellar component (unlike both Arp 220 and M82). None of the template SEDs match the steepness of the rest-frame near-IR photometry of the power-law DOGs. This may indicate that obscured AGN dominate stellar emission to a greater extent in power-law DOGs than in Mrk 231.

One striking feature of this plot is the steep slope shown by DOGs in the rest-frame 1-4m. Whereas Mrk 231, M82, and Arp 220 all exhibit a bump in the 1-2m regime, no such feature is apparent in the DOG SEDs. This could be due to the presence of an obscured AGN outshining the stellar light, in the rest-frame near-IR. This is in constrast to rest-frame UV and optical wavelengths, where HST imaging has revealed that stellar light appears to dominate (Bussmann et al. 2009). As noted previously, DOGs have far-IR to mid-IR flux density ratios more similar to Mrk 231 than Arp 220. The composite SMG template overpredicts the far-IR flux for a given mid-IR flux in all cases where we have 350m detections. We note that adding an additional warm dust component ( K; possibly powered by an AGN) to the composite SMG SED (as was done in Pope et al. 2008a) improves the quality of the fit over the rest-frame 8-100m. However, this composite SMG + AGN template retains a strong cool dust (K) component that over-predicts the amount of emission at 1mm. If this type of SED was appropriate for the power-law DOGs investigated in this paper, they would have been easily detected by CARMA.

An alternative way of displaying this information is shown in Figure 5. In each panel, the flux density ratio far-IR:mid-IR is plotted as a function of the flux density ratio mid-IR:optical (). The top two panels show on the y-axis, while the bottom two panels show on the y-axis. SMGs and various Spitzer-selected sources are shown in the plots, divided into those that are detected in the (sub-)mm on the left and those that are not detected on the right. The SMG, XFLS, and SWIRE -band data come from Dye et al. (2008), Yan et al. (2007), and Lonsdale et al. (2009), respectively. For SMGs without detections at 1200m, is estimated using the 850m flux density and the dust temperature from Coppin et al. (2008), and is represented by a red cross symbol. Dotted, dashed, and dot-dashed lines indicate the evolution of Mrk 231, Arp 220, and M82, respectively, on this diagram over redshifts of . Compared to SMGs, DOGs in this sample have redder flux density ratios in the mid-IR:optical but bluer far-IR:mid-IR ratios. This cannot be explained by an enhancement of the 24m flux due to PAH emission, since mid-IR spectra of these DOGs show power-law continua with silicate absorption and weak or absent PAH emission features (Houck et al. 2005). Instead, the most likely explanation is that obscured AGN emission boosts the mid-IR continuum (e.g., Rieke & Lebofsky 1981) relative to both the optical and far-IR.

Figure 5.— Top and bottom panels show 350m/24m and 1200m/24m flux density ratios, respectively, as a function of -[24] (Vega mag) color. For clarity, sources are separated into those detected at 350m or 1200m on the left and those that are not detected on the right. Objects that qualify as a DOG ( - [24] 14) are shown with a filled symbol. Black circles indicate DOGs in Boötes. Red squares show measurements of SMGs (Coppin et al. 2008), while red crosses show predicted values based on 850m photometry (see text for details). Green stars show Spitzer-selected bump sources from SWIRE (Lonsdale et al. 2009). Orange triangles (teal triangles) show similarly identified sources from the XFLS dominated by PAH emission (silicate absorption) features (Sajina et al. 2008). Magenta inverted triangles show XFLS sources from Sajina et al. (2008). Finally, dotted, dashed, and dot-dashed lines show the evolution of Mrk 231, Arp 220, and M82 in this parameter space over redshift 1 to 3. The DOGs studied in this paper have some of the reddest colors and lowest far-IR/mid-IR flux density ratios of other ULIRGs like SMGs or other Spitzer-selected sources.

3.2. IR Luminosities

In this section, we provide the best available estimates of the total IR luminosities (; 8-1000m rest-frame) of the sample in this paper based on the 350m imaging. We then compare these with estimates based solely on the 24m flux density, and also with estimates of the far-IR luminosity (; 40-500m, rest-frame) based on a modified black-body which has been scaled to match the sub-mm photometry.

The qualitative SED comparison from section 3.1 suggests that Mrk 231 provides a reasonable fit to the far-IR photometry. In addition, analysis of 70m and 160m photometry of a sample of these types of AGN-dominated DOGs has suggested that Mrk 231 provides a reasonable approximation of the full SED (Tyler et al. 2009, see table 3 for overlap between that study and this one). Therefore, as the best measure of , we integrate (over 8-1000m rest-frame) a redshifted Mrk 231 template which has been scaled to match the observed 350m flux density (or 3- limit, in the case of a non-detection). These values are tabulated in the first column of Table 4. For sources with detections at 350m, is in the range of , with a median value of .

The second column of Table 4 shows the rest-frame 8m luminosity, , while the third column shows the total IR luminosity (8-1000m, rest-frame) based on the spectroscopic redshift and an empirically determined relationship between and : (Caputi et al. 2007). This approach was used by Dey et al. (2008) in determining the contribution of DOGs to the total IR luminosity density of all galaxies. values range over (0.5-8.1)  10, with a median value of 2.3  10. The Caputi et al. derived value for is consistent with the 350m-based estimate (or 3- limit, in the case of non-detections) in 6/12 of the power-law DOGs studied here. In the remaining half of the sample, the Caputi et al. relation overestimates in 5/6 targets. In only one DOG (S8) is the 350m emission brighter than would be expected based on the 24m flux density, redshift, and the Caputi et al. relation. This implies that measurements of the IR luminosity density of DOGs relying solely on the 24m flux density will tend to overestimate their true contribution, consistent with what has been found in a recent study of faint (Jy) DOGs in GOODS-N (Pope et al. 2008a). Quantifying the extent of this effect will require much larger samples of DOGs with sub-mm measurements, the kind that will result from wide field surveys with Herschel and SCUBA-2.

Finally, the last column of Table 4 shows FIR luminosities computed from the integral over 40-500m (rest-frame) of the best-fit modified black-body (described in more detail in section 3.3). These values are tabulated only for those sources with CARMA 1mm imaging. We find values of , implying . In contrast, Mrk 231 has , underscoring the fact that the IR luminosity of these DOGs is dominated by mid-IR emission rather than FIR emission.

aaIntegral over 8-1000m of redshifted Mrk 231 template normalized at 350m. bbEstimated from  (8m) - relation from Caputi et al. (2007). ddUncertainties shown reflect 350m photometric uncertainties. Addtional systematic uncertainties associated with the adoption of a Mrk 231 template are not included.
Source Name () () () ()
S1 ee3 upper limits. 2.2 23
S2 ddUncertainties shown reflect 350m photometric uncertainties. Addtional systematic uncertainties associated with the adoption of a Mrk 231 template are not included. 2.4 25 9.1
S3 ddUncertainties shown reflect 350m photometric uncertainties. Addtional systematic uncertainties associated with the adoption of a Mrk 231 template are not included. 7.2 81 10
S4 ee3 upper limits. 5.2 57
S5 ee3 upper limits. 53 5.1
S6 ee3 upper limits. 1.5 15
S7 ddUncertainties shown reflect 350m photometric uncertainties. Addtional systematic uncertainties associated with the adoption of a Mrk 231 template are not included. 2.1 22
S8 ddUncertainties shown reflect 350m photometric uncertainties. Addtional systematic uncertainties associated with the adoption of a Mrk 231 template are not included. 1.1 11
S9 ee3 upper limits. 2.9 31
S10 ee3 upper limits. 6.2 69
S11 ee3 upper limits. 3.9 42
S12 ee3 upper limits. 2.0 21
Table 4Luminosities

3.3. Constraints on Dust Properties

Additional constraints can be placed on the nature of the cold dust emission from the two sources (S2 and S3) with CARMA. The 1mm non-detections imply warm dust temperatures. If we compute the predicted flux density at 1mm based on the observed 24m flux density and assuming the three local galaxy SED templates (M82, Arp 220, and Mrk 231) used in the previous sections, we find values of 4.4, 740, 1.8 mJy, and 5.8, 120, 2.7 mJy for the two sources respectively. For S2, this implies that our 3- limit (mJy) is close to the level we would expect if Mrk 231 is an appropriate SED, while the other two SEDs are clearly inconsistent with the data. For S3, the limit (mJy) is inconsistent with each of the SEDs, implying a warmer dust temperature than even Mrk 231.

A more quantitative approach is to minimize the residuals between the 160m, 350m, and 1mm data and that expected from a modified black body. Doing this, we can constrain the dust temperature, , and the dust emissivity index, , of each target. The best-fit quantities and their uncertainties are estimated using a bootstrap technique that mimics the procedure used by Dunne et al. (2000). Briefly, for each flux density measurement, a set of 100 artificial flux densities are generated using a Gaussian random number generator. The mean value and dispersion of the distribution of artificial flux densities are set by the measurement value and its 1 uncertainty, respectively (for non-detections, we assume a mean value of 0 and force artifical flux densities to be positive). Each artificial SED is used to construct a distribution of best-fit scaling factors and associated values for a modified black-body with a given combination of and .

Figure 6 shows the median contours for the grid of and values we have sampled in the model fitting for each source observed by CARMA. The contours show the degeneracy between and , and illustrate why a perfect fit is not possible in spite of the fact that a model with three parameters is being fit to three data points (i.e., the input parameters are not fully independent). The distribution of found for SMGs based on 350m imaging from Coppin et al. (2008) and Kovács et al. (2006) is displayed in the lower panel. Two-sided Kolmogorov-Smirnov tests suggest that the distribution of (assuming ) for each CARMA target is highly unlikely to be drawn from the same parent distribution as the combined sample of SMGs (4% and 0.008%, for the two CARMA targets respectively). In both of the CARMA targets, warmer dust temperatures are required due to the non-detection at 1mm. The data cannot rule out models with higher values of , but models of dust grains as well as Galactic and extra-galactic observations consistently suggest (Hildebrand 1983; Dunne & Eales 2001). For , we reject K and 45K at the 95% confidence level, in S2 and S3 respectively. For , we can reject  K and 37 K.

Figure 6.— contours based on modified black-body fits to sources with CARMA data. Lines indicate 80%, 95%, and 99.9% confidence intervals. Also shown are distributions for SMGs with 350m data (Kovács et al. 2006; Coppin et al. 2008). The 95% confidence levels for S2 and S3 suggest dust temperature limits that would place them in the warmest 50% and 15%, respectively, of SMGs.

3.4. Dust Masses

Assuming optically thin sub-mm emission, cold dust masses can be estimated from the 350m photometry (Hughes et al. 1997):


where is the observed 350m flux density, and and are, respectively, the values of the mass absorption coefficient and black-body function at the rest frequency and dust temperature . The appropriate value for is uncertain to at least a factor of two (Dunne et al. 2003); we use a value interpolated from Draine (2003) (cmg).

The results from section 3.3 suggest that  K for two of the sources. Adopting the average of these limits (K) for all of the sources in this sample, then dust mass limits are in the range ( 10 (median value of 5.110) for the five objects with detections at 350m. The 3 upper limits on the dust masses of the remaining sample range from ( 10. Warmer values of would lead to smaller inferred dust masses (e.g., increasing the dust temperature by 10 K implies 50% lower dust masses). The dust masses are presented in Table 5.

These values agree with those of Bussmann et al. (2009), where dust masses of a sample of 31 power-law dominated DOGs were estimated using predicted 850m flux densities based on Mrk 231 templates and the measured 24m flux density. In previous work, we assumed a of 75 K and found a median dust mass of 1.610. This is consistent with the notion that Mrk 231 accurately characterizes the far-IR SED of power-law DOGs, as described in section 3.1 and in Tyler et al. (2009).

Finally, assuming a gas mass to dust mass ratio of 120 (as was found in a study of the nuclear regions of nearby LIRGs; see Wilson et al. 2008), then the gas masses can be estimated as well. Using the assumed ratio, we find gas masses of ( 10 (median value of 610) for the detected objects and gas mass 3 limits of ( 10 in the remaining sample. We caution that this is very uncertain; Kovács et al. (2006) report a gas mass to dust mass ratio of for SMGs, assuming cmg. If this gas to dust mass ratio is appropriate for our sample, then the implied gas masses will be a factor of two lower ()

3.5. Stellar Masses

In this section, we describe the methodology and present estimates for the stellar masses of the DOGs in this sample.

3.5.1 Methodology

To estimate stellar masses, we rely on Simple Stellar Population (SSP) template SEDs from the Bruzual & Charlot (2003) population synthesis library. All models used here have ages spaced logarithmically from 10 Myr up to 1 Gyr, solar metallicity, a Chabrier initial mass function (IMF) over the mass range (Chabrier 2003), and use the Padova 1994 evolutionary tracks (Girardi et al. 1996). The reddening law from Calzetti et al. (2000) is used between m and that of Draine (2003) for longer wavelengths. This method is similar to that used in Bussmann et al. (2009).

For sources at whose mid-IR luminosity is dominated by stellar light, IRAC photometry samples the SED over the wavelength range where emission from asymptotic and red giant branch stars as well as low-mass main-sequence stars produces an emission peak at rest-frame 1.6m. In such cases, for given assumptions regarding the star-formation history, metallicity, and IMF, stellar mass estimates can be obtained via stellar population synthesis modeling. One goal of this work is to estimate stellar masses using self-consistent modeling of photometry measured at similar wavelengths for a variety of dusty galaxies. Therefore, we apply this method to determine stellar masses in SMGs as well as XFLS and SWIRE sources. The IRAC data for each of these galaxy populations comes from, respectively, Dye et al. (2008), Lacy et al. (2005), and Lonsdale et al. (2009).

The DOGs studied in this paper have mid-IR SEDs that are dominated by a power-law component, suggesting that obscured AGN emission is overwhelming the stellar flux at these wavelengths. The shape of the mid-IR SED therefore provides limited constraints on the stellar population and additional information is needed to estimate the stellar mass of these sources. To overcome this challenge, SSP models were fit to high spatial-resolution HST photometry in the rest-frame UV (WFPC2/F606W or ACS/F814W) and rest-frame optical (NIC2/F160W) from Bussmann et al. (2009). Two sources currently lack HST data (SST24 J142827.2+354127 and SST24 J143411.0+331733333 HST data exist for this source but will be presented in a separate paper (Prescott et al., in prep.)) and so are excluded from this analysis.

In principle, rest-frame near-IR data offer a means to estimate the SSP age and independently, since rest-frame optical and near-IR photometry sample the SED above the 4000 Å break while the rest-frame UV photometry samples galaxy light below the 4000 Å break. However, these data come from the IRAC 3.6m images of the Boötes field (Spitzer Deep Wide-Field Survey; Ashby et al., in prep.), where the spatial resolution is insufficient to resolve the nuclear source from the extended galaxy component. In this case, there are only limited constraints on the amount of non-stellar (i.e., obscured AGN) emission at 3.6m. We have explored the effect of this uncertainty on the fitting process by artificially reducing the 3.6m flux by 50% (corresponding to the situation where the 3.6m emission is equal parts starlight and AGN) and re-analyzing the data. Comparing results, we find that higher AGN fractions imply younger inferred SSP ages and higher values. Because the AGN fraction in these sources is currently unknown at 3.6m, we are unable to constrain both the age and independently.

Although the usage of photometry at different wavelengths is not ideal for the purposes of comparing stellar masses between galaxy populations, this method remains valuable because the models being fit to the data are the same for each galaxy population. Indeed, recent work has suggested that the dominant source of systematic uncertainty in stellar mass estimates of -selected galaxies at is the use of different stellar population synthesis codes (Muzzin et al. 2009), and that these systematics often dominate the formal random uncertainty. As long as the parameters of the model used here (such as the IMF, star-formation history, metallicity, etc.) do not vary from population to population, then the comparison presented here should be valid in a global sense.

3.5.2 Stellar Mass Estimates

Figure 7 shows contours for a grid of SSP ages and values. The contours trace lines of 80%, 95%, and 99.9% confidence intervals allowed by the photometric uncertainties (estimated using a bootstrap method similar to that outlined in section 3.3). The solid grey lines trace iso-mass contours and show the range in stellar masses allowed by the photometric uncertainties. The best-fit SSP model parameters ( and ) are printed in each panel and shown in Table 5.

Figure 7.— contours based on SSP fits to HST imaging in -band and - or -band. Black lines indicate 80%, 95%, and 99.9% confidence intervals. Grey contours trace lines of constant stellar mass for the given HST photometry, starting with in the bottom left and increasing by 0.5 dex towards the upper right. In the top right corner of each panel is the minimum value and the associated stellar mass, in units of log(). The bottom left corner contains the source identifier. Not shown are sources S2 and S7, since these targets have no HST imaging available. The best-fit stellar masses range from .

The stellar masses in the sample range from , with a median value of . The values range from , with a median value of 0.69. In one case (SST24 J143001.9+334538, or S3), the photometric uncertainty is so large that over the full range of and SSP age that we have sampled and so the range of acceptable fits is very large. For this source, we quote the 3 upper limit on the stellar mass based on the photometric uncertainty.

The ratio of the stellar to gas mass, , is a measure of the evolutionary state of the galaxy, with larger values indicating more processing of gas to stars. Our estimates of , computed assuming , are presented in Table 5. We caution that the gas mass to dust mass ratio is highly uncertain. In SMGs, there is evidence suggesting it is (Kovács et al. 2006). Adopting this lower value would imply lower gas masses by a factor of two and hence double our estimates.

Clustering studies suggest that the most luminous DOGs reside in very massive halos ( Brodwin et al. 2008). It is tempting to attribute the low stellar masses we estimate for DOGs to youth. However, the absolute stellar masses we compute are extremely uncertain. For example, the use of a Salpeter IMF rather than a Chabrier IMF would approximately double our stellar mass estimates (Bruzual & Charlot 2003). Beyond the choice of what IMF slope to use, the mass-to-light ratio of a model galaxy (for a given rest-frame near-UV - color) can vary significantly depending on the details of its star-formation history, the clumpiness of its interstellar medium and the associated dust attenuation law, as well as how advanced stages of stellar evolution are treated, such as blue stragglers, thermally-pulsating asymptotic giant branch stars, etc. (Conroy et al. 2009). In light of these uncertainties, the fact that our stellar mass estimates are low () compared to the dark matter haloes in which we believe they reside is not yet a cause for concern – a quantitative study of the maximum possible stellar mass allowed by the photometry (by examining results from different stellar population synthesis codes, star-formation histories, metallicities, etc.) would be the best way to approach this issue in the near-term, but is beyond the scope of the current work.

aaDust mass assuming  K. bbStellar mass estimated from fitting photometry in the rest-frame UV and optical ( and respectively). Range given reflects 95% confidence intervals based only on photometric uncertainty.
Source Name () ()
Table 5Dust Masses and Stellar Properties

4. Discussion

In this section, we seek to understand the role of DOGs in galaxy evolution and their relation to other high- galaxy populations. We begin by motivating a comparison sample of such objects, including SMGs and Spitzer-selected ULIRGs from the XFLS and SWIRE survey. We then examine how the measured properties differ from population to population. We end with the implications of these comparisons for models of galaxy evolution.

4.1. Related Galaxy Populations

4.1.1 SMGs

SMGs represent an interesting population of galaxies for comparison with DOGs because they are selected at sub-mm wavelengths where the dominant emission component is cold dust (K). In contrast, DOGs are selected predominantly by their brightness at 24m and therefore should be dominated by hot dust. Despite this fundamental distinction, these two galaxy populations have similar number densities and redshift distributions (Chapman et al. 2005; Dey et al. 2008; Blain et al. 2004). Recent evidence suggests that 24m-faint ( mJy) DOGs have a composite SED whose shape in the far-IR closely mimics that of the average bright (mJy) SMG (Pope et al. 2008a). Futhermore, 24m-faint DOGs and SMGs have similar real space correlation lengths (Mpc), yet there is tentative evidence that DOG clustering strength increases with 24m flux density (Mpc for DOGs with  mJy; Brodwin et al. 2008). While these results are suggestive of an association between the two populations, the details of such a connection are not yet clear. In an effort to study this connection via their far-IR properties, we will compare the data presented in this paper with SHARC-II 350m and MAMBO 1.2mm imaging of 25 SMGs from the Submillimetre Common User Bolometer Array (SCUBA) HAlf Degree Extragalactic Survey (Laurent et al. 2006; Kovács et al. 2006; Coppin et al. 2008; Greve et al. 2004).

4.1.2 XFLS Sources

A set of Spitzer-selected galaxies from the 4 deg XFLS share many properties with the 24m-bright DOGs (Yan et al. 2007). The specific selection criteria are similar, although not necessarily as extreme in their IR-optical flux density ratios:  mJy, , and (in comparison, DOGs have ). Spitzer/IRS spectroscopy of these objects has revealed strong silicate absorption and in some cases PAH emission features on par with those of SMGs (Sajina et al. 2007). This suggests that the XFLS sources are composite AGN/starburst systems and may represent a transition phase between (un)obscured quasars and SMGs (Sajina et al. 2008). MAMBO 1.2mm observations of 44 XFLS sources have allowed a detailed study of their far-IR properties and have suggested 710 (Sajina et al. 2008).

4.1.3 SWIRE Sources

The last set of comparison galaxies we consider are Spitzer-selected sources from the SWIRE survey (Lonsdale et al. 2009). Like DOGs in Boötes and the XFLS sources, they have large IR-optical flux density ratios. However, an additional criterion has been applied to identify sources with significant emission at rest-frame 1.6m due to evolved stellar populations. For sources at , this means selecting objects whose mid-IR spectrum peaks at 5.8m. Although spectroscopic redshifts are not available for most of this sample, SED fitting has suggested photometric redshifts consistent with and stellar masses of (0.2 - 6)10 (Lonsdale et al. 2009). MAMBO 1.2mm photometry for 61 of these SWIRE sources has indicated far-IR luminosities of 10-10 (Lonsdale et al. 2009).

4.2. Comparison of Measured Properties

Our results from section 3 represent our best estimates of , , , , and for the DOGs in the sample. In Table 6, we give the median value of these quantities for DOGs in Boötes (from this paper), SMGs, and XFLS and SWIRE sources. In computing these median values, we do not consider sources at ; nor do we consider sources without detections at (sub-)mm wavelengths (see discussion on caveats to the analysis at the end of this section). Table 6 also makes a distinction between XFLS sources whose mid-IR spectra are dominated by strong PAH features (XFLS PAH) and those that show weak or absent PAH features (XFLS weak-PAH). Each of these galaxy populations is further subdivided into those that qualify as DOGs () and those that do not.

Source -[24] () () (K) () () aaComputed using .
BoötesbbIncludes only DOGs detected at 350m. 14 5 23 10 45 0.5
XFLSccIntegral over 40-500m of best-fit modified black-body (only sources with CARMA 1mm data). ALL 11 7.7 2.7 32 5.0 13 0.48
14 6 8.7 1.1 27 7.3 10 0.21
14 5 6.5 4.6 37 2.3 16 0.58
XFLS PAHccXFLS sources with MAMBO 1.2mm detections (Sajina et al. 2008). ALL 5 5.7 1.8 31 4.6 22 0.54
14 2 6.3 1.6 29 7.5 23 0.30
14 3 5.4 2.0 32 2.8 22 0.69
XFLS weak-PAHccXFLS sources with MAMBO 1.2mm detections (Sajina et al. 2008). ALL 6 9.4 3.4 32 5.3 5.0 0.25
14 4 3.5 0.8 26 7.2 3.8 0.17
14 2 14 8.0 45 1.6 7.3 0.40
SWIREddSWIRE sources with MAMBO 1.2mm detections (Lonsdale et al. 2009). ALL 19 6.5 3.1 32 6.7 28 0.53
14 16 6.7 3.1 32 6.1 28 0.56
14 3 5.5 2.6 30 9.9 24 0.35
SMGeeFrom compilation of Coppin et al. (2008). ALL 18 6.9 3.2 35 1.5 10 0.60
14 4 7.8 3.6 28 2.9 6.9 0.28
14 14 6.1 2.9 37 1.1 12 0.71
Table 6Average High- Galaxy Properties

The primary feature of this comparison is that the relative uncertainty in the estimated parameters between galaxy populations has been reduced by computing the respective values self-consistently with the methods outlined in section 3. The exceptions to this rule are and (note that while the photometry used to determine for Boötes DOGs is different than for the other galaxy populations, the methodology used is the same, including the use of the same set of model SSP templates). Our method of computing relies on the assumption that Mrk 231 represents a reasonable approximation of the source SED. For many SMGs as well as XFLS and SWIRE sources, this is an unrealistic assumption. Instead, we estimate from , assuming (1) ; (2) ; (3) , where is a factor , depending on and (Helou et al. 1988) and is the typical AGN fraction of the galaxy population. For SMGs, we adopt the conservative upper limit from Pope et al. (2008b) of . For SWIRE sources and XFLS PAH sources, this fraction is (Pope et al. 2008b; Lonsdale et al. 2009; Sajina et al. 2008), while for XFLS weak-PAH sources we use 0.7 (Sajina et al. 2008).

The values given in the literature are adopted for each source. It should be noted that for SWIRE sources are uncertain due to the lack of data near the far-IR peak (i.e., observed-frame 160 or 350 m). Lonsdale et al. (2009) analyze the stacked signal at 160m from these sources and find that the is higher by as much as 10 K than what is assumed in their Table 6. For a given 1.2mm flux, increasing by 10 K will increase by a factor of 3 and decrease by 50%.

The key result from Table 6 is that while the DOGs in our sample have lower dust masses than the other galaxy populations by a factor of , they have higher total IR and far-IR luminosities by factors of 2. This distinction is driven by the difference in , as DOGs in the sample have higher values by 10-20 K compared to the other galaxy populations.

In terms of the stellar and gas mass estimates, the relationship between DOGs in Boötes (i.e., the sample studied in this paper) and the remaining galaxy populations is unclear. Even if a single dust temperature and a single dust to gas mass ratio for each of the sources studied in this paper is adopted, the uncertainties on the stellar mass estimates are large enough to allow greatly varying stellar mass to gas mass ratios. Sources satisfying (i.e., DOGs) tend to have higher gas masses compared to sources (under the assumption of a constant dust-to-gas mass ratio). This difference is at least in part due to a difference in dust temperatures; within this sample, DOGs have lower dust temperatures than non-DOGs. This is in contrast with the evidence for high dust temperatures in the Boötes DOGs studied in this paper and may be an indication that mm-detected DOGs represent a special subset of DOGs that is more representative of the mm-selected galaxy population than the DOG population.

An important caveat to this comparison is that we are dealing with small sample sizes due to incomplete coverage at one or more bands from the mid- to the far-IR. For instance, while every DOG has a measured 24m flux density, very few have been observed at 350m, and only two have been observed at 1mm. Similarly, few XFLS and SWIRE sources have been detected at 1mm and even fewer have been observed at 350m. Although SMGs are the best-studied class of objects within this set of populations, they too suffer from low-number statistics. Larger sample sizes in the critical 200-500m regime will arrive following the analysis of wide-field survey data from the Balloon-borne Large Aperture Submillimeter Telescope (e.g., Pascale et al. 2009) and the Herschel Space Observatory.

4.3. Implications for Models of Galaxy Evolution

One of the major open questions in galaxy evolution is the effect that AGN have on their host galaxies. In the local universe, there is observational evidence that ULIRGs dominated by warm dust serve as a transition phase between cold dust ULIRGs and optically luminous quasars and that this transition may be driven by a major merger (Sanders et al. 1988a, b). Recent theoretical models of quasar evolution based on numerical simulations of major mergers between gas-rich spirals have suggested that the subsequent growth of a super-massive black hole can regulate star-formation via a feedback effect which re-injects energy into the interstellar medium (ISM) and expels the remaining gas that would otherwise form stars (Hopkins et al. 2006).

Although the notion that local ULIRGs are associated with mergers is well accepted (e.g., Sanders & Mirabel 1996), the picture is less clear at high redshift. Morphological studies of high- galaxies suffer from surface brightness dimming, making the detection of faint merger remnant signatures difficult (e.g., Dasyra et al. 2008; Melbourne et al. 2008; Bussmann et al. 2009; Melbourne et al. 2009). However recent theoretical work on the cosmological role of mergers in the formation of quasars and spheroid galaxies suggests that they dominate the quasar luminosity density compared to secular processes such as bars or disk instabilities (Hopkins et al. 2008).

If major mergers drive the formation of massive galaxies at high redshift, then one possible interpretation of our results involves an evolutionary scenario in which these sources represent a very brief but luminous episode of extreme AGN growth just prior to the quenching of star formation. In such a scenario, SMGs and the brightest 24m-selected sources represent the beginning and end stages, respectively, of the high star-formation rate, high IR luminosity phase in massive galaxy evolution. Consistent with this scenario is that we find 24m-bright DOGs in Boötes to have higher dust temperatures (possibly from AGN heating of the dust) than SMGs and less extreme Spitzer-selected sources. We caution, however, that these results are consistent with any evolutionary model in which the 24m-bright phase follows the sub-mm bright phase, be it driven by major mergers, minor mergers, or some secular process.

Finally, we stress that larger samples of mm and sub-mm imaging of Spitzer-selected galaxies are needed in order to understand their role in galaxy evolution fully by comparing samples of similar number density, clustering properties, etc. Much of this will be provided by upcoming Herschel and SCUBA-2 surveys. In the more immediate future, 1mm imaging with currently available instruments such as the Astronomical Thermal Emission Camera (AzTEC) and the MAx-Planck Millimeter BOlometer Array (MAMBO) will be critical to constraining the cold dust properties of Spitzer-selected galaxies. Only when these surveys have obtained statistically significant numbers of detections or stringent upper limits will we be able to make definitive conclusions regarding the nature of the link between AGN and starbursts in the formation of the most massive galaxies.

5. Conclusions

We present CSO/SHARC-II 350m and CARMA 1mm photometry of DOGs in the Boötes Field. The major results and conclusions from this study are the following:

  1. At 350m, 4/5 DOGs are detected in data with low rms levels (mJy) and 0/8 DOGs are detected in data with medium to high rms levels (mJy). At 1mm, a subset of two DOGs were observed but not detected.

  2. Mrk 231 is confirmed as a valid template for the SEDs of the DOGs in this sample. This suggests the 24m bright ( mJy) population of DOGs is dominated by warm dust, possibly heated by an AGN. Cold dust templates such as Arp 220 are inconsistent with the data in all twelve objects studied.

  3. Trends in the flux density ratios 350m/24m and 1200m/24m with the -[24] color () show that DOGs in this sample have elevated 24m emission relative to SMGs, most likely due to an obscured AGN.

  4. The non-detections at 1mm imply greater than 35-60 K for two objects.

  5. If the dust properties of the two DOGs observed at 1mm apply generally to the 24m bright DOGs, then we estimate dust masses for these sources of . Lower would imply higher dust masses and vice versa.

  6. In comparison to other ULIRGs, DOGs have warmer dust temperatures that imply higher IR luminosities and lower dust masses. This may be an indication that AGN growth has heated the ambient ISM in these sources.

  7. Our stellar mass estimates provide weak evidence indicating that the 24m-bright DOGs may have converted more gas into stars than SMGs or other Spitzer-selected sources, consistent with them representing a subsequent phase of evolution. An important caveat to this conclusion is that we have assumed DOGs and SMGs share the same gas mass to dust mass ratio. Testing this assumption will require new data and will be an important goal of future work.

This work is based in part on observations made with the Spitzer Space Telescope, which is operated by the Jet Propulsion Laboratory, California Institute of Technology under NASA contract 1407. Spitzer/MIPS guaranteed time observing was used to image the Boötes field at 24m and is critical for the selection of DOGs. We thank the SDWFS team (particularly Daniel Stern and Matt Ashby) for making the IRAC source catalogs publicly available. Data from the original IRAC shallow survey were used for intial stellar mass estimates. We thank the anonymous referee for a thorough review of the manuscript that helped improve the paper.

We are grateful to the expert assistance of the staff of Kitt Peak National Observatory where the Boötes field observations of the NDWFS were obtained. The authors thank NOAO for supporting the NOAO Deep Wide-Field Survey. In particular, we thank Jenna Claver, Lindsey Davis, Alyson Ford, Emma Hogan, Tod Lauer, Lissa Miller, Erin Ryan, Glenn Tiede and Frank Valdes for their able assistance with the NDWFS data. We also thank the staff of the W. M. Keck Observatory, where some of the galaxy redshifts were obtained.

RSB gratefully acknowledges financial assistance from HST grant GO10890, without which this research would not have been possible. Support for Program number HST-GO10890 was provided by NASA through a grant from the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy, Incorporated, under NASA contract NAS5-26555. The research activities of AD are supported by NOAO, which is operated by the Association of Universities for Research in Astronomy (AURA) under a cooperative agreement with the National Science Foundation. Support for E. Le Floc’h was provided by NASA through the Spitzer Space Telescope Fellowship Program.

Facilities used: Spitzer, CSO, and CARMA. This research made use of CSO (SHARC-II) and CARMA data. Support for CARMA construction was derived from the states of California, Illinois, and Maryland, the Gordon and Betty Moore Foundation, the Kenneth T. and Eileen L. Norris Foundation, the Associates of the California Institute of Technology, and the National Science Foundation. Ongoing CARMA development and operations are supported by the National Science Foundation under a cooperative agreement, and by the CARMA partner universities.


  • Ashby et al. (2009) Ashby, M. L. N., et al. 2009, ApJ, 701, 428
  • Blain et al. (2004) Blain, A. W., Chapman, S. C., Smail, I., & Ivison, R. 2004, ApJ, 611, 725
  • Brand et al. (2008) Brand, K., Weedman, D. W., Desai, V., Le Floc’h, E., Armus, L., Dey, A., Houck, J. R., Jannuzi, B. T., Smith, H. A., & Soifer, B. T. 2008, ApJ, 680, 119
  • Brodwin et al. (2008) Brodwin, M., Dey, A., Brown, M. J. I., Pope, A., Armus, L., Bussmann, S., Desai, V., Jannuzi, B. T., & Le Floc’h, E. 2008, ApJ, 687, L65
  • Bruzual & Charlot (2003) Bruzual, G. & Charlot, S. 2003, MNRAS, 344, 1000
  • Bussmann et al. (2009) Bussmann, R. S., et al., 2009, ApJ, 693, 750
  • Calzetti et al. (2000) Calzetti, D., Armus, L., Bohlin, R. C., Kinney, A. L., Koornneef, J., & Storchi-Bergmann, T. 2000, ApJ, 533, 682
  • Caputi et al. (2007) Caputi, K. I., Lagache, G., Yan, L., Dole, H., Bavouzet, N., Le Floc’h, E., Choi, P. I., Helou, G., & Reddy, N. 2007, ApJ, 660, 97
  • Chabrier (2003) Chabrier, G. 2003, PASP, 115, 763
  • Chapman et al. (2005) Chapman, S. C., Blain, A. W., Smail, I., & Ivison, R. J. 2005, ApJ, 622, 772
  • Conroy et al. (2009) Conroy, C., White, M., & Gunn, J. E. 2009, ArXiv e-prints
  • Coppin et al. (2005) Coppin, K., Halpern, M., Scott, D., Borys, C., & Chapman, S. 2005, MNRAS, 357, 1022
  • Coppin et al. (2008) Coppin, K., et al., 2008, MNRAS, 384, 1597
  • Dasyra et al. (2008) Dasyra, K. M., Yan, L., Helou, G., Surace, J., Sajina, A., & Colbert, J. 2008, ApJ, 680, 232
  • de Vries et al. (2002) de Vries, W. H., Morganti, R., Röttgering, H. J. A., Vermeulen, R., van Breugel, W., Rengelink, R., & Jarvis, M. J. 2002, AJ, 123, 1784
  • Desai et al. (2006) Desai, V., et al., 2006, ApJ, 641, 133
  • Dey et al. (2005) Dey, A., et al., 2005, ApJ, 629, 654
  • Dey et al. (2008) Dey, A., et al., 2008, ApJ, 677, 943
  • Donley et al. (2007) Donley, J. L., Rieke, G. H., Pérez-González, P. G., Rigby, J. R., & Alonso-Herrero, A. 2007, ApJ, 660, 167
  • Draine (2003) Draine, B. T. 2003, ARA&A, 41, 241
  • Dunne et al. (2000) Dunne, L., Eales, S., Edmunds, M., Ivison, R., Alexander, P., & Clements, D. L. 2000, MNRAS, 315, 115
  • Dunne & Eales (2001) Dunne, L. & Eales, S. A. 2001, MNRAS, 327, 697
  • Dunne et al. (2003) Dunne, L., Eales, S. A., & Edmunds, M. G. 2003, MNRAS, 341, 589
  • Dye et al. (2008) Dye, S., et al., 2008, MNRAS, 386, 1107
  • Eisenhardt et al. (2004) Eisenhardt, P. R., et al. 2004, ApJS, 154, 48
  • Faber et al. (2003) Faber, S. M., et al., 2003, in Presented at the Society of Photo-Optical Instrumentation Engineers (SPIE) Conference, Vol. 4841, Instrument Design and Performance for Optical/Infrared Ground-based Telescopes. Edited by Iye, Masanori; Moorwood, Alan F. M. Proceedings of the SPIE, Volume 4841, pp. 1657-1669 (2003)., ed. M. Iye & A. F. M. Moorwood, 1657–1669
  • Fiore et al. (2008) Fiore, F., et al., 2008, ApJ, 672, 94
  • Franceschini et al. (2001) Franceschini, A., Aussel, H., Cesarsky, C. J., Elbaz, D., & Fadda, D. 2001, A&A, 378, 1
  • Gallagher et al. (2002) Gallagher, S. C., Brandt, W. N., Chartas, G., Garmire, G. P., & Sambruna, R. M. 2002, ApJ, 569, 655
  • Girardi et al. (1996) Girardi, L., Bressan, A., Chiosi, C., Bertelli, G., & Nasi, E. 1996, A&AS, 117, 113
  • Greve et al. (2004) Greve, T. R., Ivison, R. J., Bertoldi, F., Stevens, J. A., Dunlop, J. S., Lutz, D., & Carilli, C. L. 2004, MNRAS, 354, 779
  • Helou et al. (1988) Helou, G., Khan, I. R., Malek, L., & Boehmer, L. 1988, ApJS, 68, 151
  • Hildebrand (1983) Hildebrand, R. H. 1983, QJRAS, 24, 267
  • Hopkins et al. (2006) Hopkins, P. F., Hernquist, L., Cox, T. J., Di Matteo, T., Robertson, B., & Springel, V. 2006, ApJS, 163, 1
  • Hopkins et al. (2008) Hopkins, P. F., Hernquist, L., Cox, T. J., & Kereš, D. 2008, ApJS, 175, 356
  • Houck et al. (2004) Houck, J. R., et al., 2004, ApJS, 154, 18
  • Houck et al. (2005) Houck, J. R., Soifer, B. T., Weedman, D., Higdon, S. J. U., Higdon, J. L., Herter, T., Brown, M. J. I., Dey, A., Jannuzi, B. T., Le Floc’h, E., Rieke, M., Armus, L., Charmandaris, V., Brandl, B. R., & Teplitz, H. I. 2005, ApJ, 622, L105
  • Hughes et al. (1997) Hughes, D. H., Dunlop, J. S., & Rawlings, S. 1997, MNRAS, 289, 766
  • Hughes et al. (1994) Hughes, D. H., Gear, W. K., & Robson, E. I. 1994, MNRAS, 270, 641
  • Hutchings & Neff (1987) Hutchings, J. B. & Neff, S. G. 1987, AJ, 93, 14
  • Kovács (2006) Kovács, A. 2006, PhD thesis, Caltech
  • Kovács et al. (2006) Kovács, A., Chapman, S. C., Dowell, C. D., Blain, A. W., Ivison, R. J., Smail, I., & Phillips, T. G. 2006, ApJ, 650, 592
  • Lacy et al. (2005) Lacy, M., et al., 2005, ApJS, 161, 41
  • Laurent et al. (2006) Laurent, G. T., Glenn, J., Egami, E., Rieke, G. H., Ivison, R. J., Yun, M. S., Aguirre, J. E., Maloney, P. R., & Haig, D. 2006, ApJ, 643, 38
  • Le Floc’h et al. (2005) Le Floc’h, E., Papovich, C., Dole, H., Bell, E. F., Lagache, G., Rieke, G. H., Egami, E., Pérez-González, P. G., Alonso-Herrero, A., Rieke, M. J., Blaylock, M., Engelbracht, C. W., Gordon, K. D., Hines, D. C., Misselt, K. A., Morrison, J. E., & Mould, J. 2005, ApJ, 632, 169
  • Leong et al. (2006) Leong, M., Peng, R., Houde, M., Yoshida, H., Chamberlin, R., & Phillips, T. G. 2006, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, Vol. 6275, Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series
  • Lonsdale et al. (2009) Lonsdale, C. J., et al. 2009, ApJ, 692, 422
  • Melbourne et al. (2009) Melbourne, J., et al. 2009, AJ, 137, 4854
  • Melbourne et al. (2008) Melbourne, J., Desai, V., Armus, L., Dey, A., Brand, K., Thompson, D., Soifer, B. T., Matthews, K., Jannuzi, B. T., & Houck, J. R. 2008, AJ, 136, 1110
  • Muzzin et al. (2009) Muzzin, A., Marchesini, D., van Dokkum, P. G., Labbé, I., Kriek, M., & Franx, M. 2009, arXiv:0906.2012
  • Oke et al. (1995) Oke, J. B., Cohen, J. G., Carr, M., Cromer, J., Dingizian, A., Harris, F. H., Labrecque, S., Lucinio, R., Schaal, W., Epps, H., & Miller, J. 1995, PASP, 107, 375
  • Papovich et al. (2007) Papovich, C., Rudnick, G., Le Floc’h, E., van Dokkum, P. G., Rieke, G. H., Taylor, E. N., Armus, L., Gawiser, E., Huang, J., Marcillac, D., & Franx, M. 2007, ApJ, 668, 45
  • Pascale et al. (2009) Pascale, E., et al., 2009, Arxiv e-prints, 0904.1206P
  • Pérez-González et al. (2005) Pérez-González, P. G., Rieke, G. H., Egami, E., Alonso-Herrero, A., Dole, H., Papovich, C., Blaylock, M., Jones, J., Rieke, M., Rigby, J., Barmby, P., Fazio, G. G., Huang, J., & Martin, C. 2005, ApJ, 630, 82
  • Polletta et al. (2008) Polletta, M., Weedman, D., Hönig, S., Lonsdale, C. J., Smith, H. E., & Houck, J. 2008, ApJ, 675, 960
  • Pope et al. (2008a) Pope, A., Bussmann, R. S., Dey, A., Meger, N., Alexander, D. M., Brodwin, M., Chary, R.-R., Dickinson, M. E., Frayer, D. T., Greve, T. R., Huynh, M., Lin, L., Morrison, G., Scott, D., & Yan, C.-H. 2008a, ArXiv e-prints, 808
  • Pope et al. (2008b) Pope, A., Chary, R.-R., Alexander, D. M., Armus, L., Dickinson, M., Elbaz, D., Frayer, D., Scott, D., & Teplitz, H. 2008b, ApJ, 675, 1171
  • Rieke et al. (2008) Rieke, G. H., Alonso-Herrero, A., Weiner, B. J., Perez-Gonzalez, P. G., Blaylock, M., Donley, J. L., Marcillac, D., & . 2008, ArXiv e-prints
  • Rieke & Lebofsky (1981) Rieke, G. H. & Lebofsky, M. J. 1981, ApJ, 250, 87
  • Rieke et al. (2004) Rieke, G. H., et al., 2004, ApJS, 154, 25
  • Rigopoulou et al. (1996) Rigopoulou, D., Lawrence, A., & Rowan-Robinson, M. 1996, MNRAS, 278, 1049
  • Sajina et al. (2007) Sajina, A., Yan, L., Lacy, M., & Huynh, M. 2007, ApJ, 667, L17
  • Sajina et al. (2008) Sajina, A., Yan, L., Lutz, D., Steffen, A., Helou, G., Huynh, M., Frayer, D., Choi, P., Tacconi, L., & Dasyra, K. 2008, ApJ, 683, 659
  • Sanders et al. (2003) Sanders, D. B., Mazzarella, J. M., Kim, D.-C., Surace, J. A., & Soifer, B. T. 2003, AJ, 126, 1607
  • Sanders & Mirabel (1996) Sanders, D. B. & Mirabel, I. F. 1996, ARA&A, 34, 749
  • Sanders et al. (1988a) Sanders, D. B., Soifer, B. T., Elias, J. H., Madore, B. F., Matthews, K., Neugebauer, G., & Scoville, N. Z. 1988a, ApJ, 325, 74
  • Sanders et al. (1988b) Sanders, D. B., Soifer, B. T., Elias, J. H., Neugebauer, G., & Matthews, K. 1988b, ApJ, 328, L35
  • Sault et al. (1995) Sault, R. J., Teuben, P. J., & Wright, M. C. H. 1995, in Astronomical Society of the Pacific Conference Series, Vol. 77, Astronomical Data Analysis Software and Systems IV, ed. R. A. Shaw, H. E. Payne, & J. J. E. Hayes, 433–+
  • Silva et al. (1998) Silva, L., Granato, G. L., Bressan, A., & Danese, L. 1998, ApJ, 509, 103
  • Soifer et al. (1986) Soifer, B. T., Sanders, D. B., Neugebauer, G., Danielson, G. E., Lonsdale, C. J., Madore, B. F., & Persson, S. E. 1986, ApJ, 303, L41
  • Tyler et al. (2009) Tyler, K. D., Floc’h, E. L., Rieke, G. H., Dey, A., Desai, V., Brand, K., Borys, C., Jannuzi, B. T., Armus, L., Dole, H., Papovich, C., Brown, M. J. I., Blaylock, M., Higdon, S. J. U., Higdon, J. L., Charmandaris, V., Ashby, M. L. N., & Smith, H. A. 2009, ApJ, 691, 1846
  • Wang et al. (2008) Wang, R., Carilli, C. L., Wagg, J., Bertoldi, F., Walter, F., Menten, K. M., Omont, A., Cox, P., Strauss, M. A., Fan, X., Jiang, L., & Schneider, D. P. 2008, ApJ, 687, 848
  • Weedman et al. (2006a) Weedman, D., Polletta, M., Lonsdale, C. J., Wilkes, B. J., Siana, B., Houck, J. R., Surace, J., Shupe, D., Farrah, D., & Smith, H. E. 2006a, ApJ, 653, 101
  • Weedman et al. (2006b) Weedman, D. W., Soifer, B. T., Hao, L., Higdon, J. L., Higdon, S. J. U., Houck, J. R., Le Floc’h, E., Brown, M. J. I., Dey, A., Jannuzi, B. T., Rieke, M., Desai, V., Bian, C., Thompson, D., Armus, L., Teplitz, H., Eisenhardt, P., & Willner, S. P. 2006b, ApJ, 651, 101
  • Wilson et al. (2008) Wilson, C. D., Petitpas, G. R., Iono, D., Baker, A. J., Peck, A. B., Krips, M., Warren, B., Golding, J., Atkinson, A., Armus, L., Cox, T. J., Ho, P., Juvela, M., Matsushita, S., Mihos, J. C., Pihlstrom, Y., & Yun, M. S. 2008, ArXiv e-prints, 806
  • Yan et al. (2004) Yan, L., et al., 2004, ApJS, 154, 60
  • Yan et al. (2007) Yan, L., Sajina, A., Fadda, D., Choi, P., Armus, L., Helou, G., Teplitz, H., Frayer, D., & Surace, J. 2007, ApJ, 658, 778
  • Yang & Phillips (2007) Yang, M., & Phillips, T. 2007, ApJ, 662, 284
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
Add comment
Loading ...
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description