Extrapolation of Galactic Dust Emission at 100 Microns to Cosmic Microwave Background Radiation Frequencies Using FIRAS
Total Page:16
File Type:pdf, Size:1020Kb
Extrapolation of Galactic Dust Emission at 100 Microns to Cosmic Microwave Background Radiation Frequencies Using FIRAS The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters Citation Finkbeiner, Douglas P., Marc Davis, and David J. Schlegel. 1999. “Extrapolation of Galactic Dust Emission at 100 Microns to Cosmic Microwave Background Radiation Frequencies Using FIRAS.” The Astrophysical Journal 524 (2) (October 20): 867–886. doi:10.1086/307852. Published Version 10.1086/307852 Citable link http://nrs.harvard.edu/urn-3:HUL.InstRepos:33462888 Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of- use#LAA THE ASTROPHYSICAL JOURNAL, 524:867È886, 1999 October 20 ( 1999. The American Astronomical Society. All rights reserved. Printed in U.S.A. EXTRAPOLATION OF GALACTIC DUST EMISSION AT 100 MICRONS TO COSMIC MICROWAVE BACKGROUND RADIATION FREQUENCIES USING FIRAS DOUGLAS P. FINKBEINER AND MARC DAVIS University of California at Berkeley, Departments of Physics and Astronomy, 601 Campbell Hall, Berkeley, CA 94720; dÐnk=astro.berkeley.edu, marc=deep.berkeley.edu AND DAVID J. SCHLEGEL Princeton University, Department of Astrophysics, Peyton Hall, Princeton, NJ 08544; schlegel=astro.princeton.edu Received 1999 March 5; accepted 1999 June 8 ABSTRACT We present predicted full-sky maps of submillimeter and microwave emission from the di†use inter- stellar dust in the Galaxy. These maps are extrapolated from the 100 km emission and 100/240 km Ñux ratio maps that Schlegel, Finkbeiner, & Davis generated from IRAS and COBE/DIRBE data. Results are presented for a number of physically plausible emissivity models. The correlation of COBE/FIRAS data with the simple Schlegel, Finkbeiner, & Davis (l2 emissivity power law) extrapolation is much tighter than with other common dust templates such as H I column density or 100 km emission. Despite the apparent success of the Schlegel, Finkbeiner, & Davis extrapolation, the assumed l2 emissivity is inconsistent with the FIRAS data below 800 GHz. Indeed, no power-law emissivity function Ðts the FIRAS data from 200 to 2100 GHz. In this paper we provide a formalism for a multicomponent model for the dust emission. A two-component model with a mixture of ““ silicate ÏÏ and ““ carbon-dominated ÏÏ grains (motivated by Pollack et al.) provides a Ðt to an accuracy of D15% to all the FIRAS data over the entire high-latitude sky. Small systematic di†erences are found between the atomic and molecular phases of the ISM. COBE/DMR has observed microwave emission that is correlated with thermal dust emission. However, this emission is higher than our model predicts by factors of 1.2, 2.4, and 20 at 90, 53, and 31 GHz, respectively. This provides evidence that another emission mechanism dominates dust emission at frequencies below D60 GHz. Our predictions for the thermal (vibrational) emission from Galactic dust at l\3000 GHz are available for general use. These full-sky predictions can be made at the DIRBE resolution of 40@ or at the higher resolution of6.1@ from the Schlegel, Finkbeiner, & Davis DIRBE-corrected IRAS maps. Subject headings: dust, extinction È infrared: ISM: continuum È submillimeter INTRODUCTION 1. In this paper, we consider the use of the SFD98 dust map The pioneering Infrared Astronomy Satellite (IRAS) led as a predictor for microwave emission from Galactic dust. to the discovery of the ubiquitous infrared cirrus, whose The SFD98 map is based solely upon 100È240 km (1250È thermal emission is especially visible in the 100 km band 3000 GHz) emission. Extrapolation to microwave fre- (Low et al. 1984). This cirrus, with a characteristic tem- quencies is very sensitive to the details of the composition perature of D20 K, arches across the sky in long Ðlamen- and emissivity properties of the dust. We show that the l2 tary chains and is present at all Galactic latitudes. However, emissivity assumed by SFD98 is inconsistent with the 100È IRAS was optimized for the detection of point sources, and 2100 GHz emission probed by the COBE Far Infrared its ability to map the di†use cirrus was less than optimal. Absolute Spectrophotometer (FIRAS). We use these FIRAS Because of calibration drifts and hysteresis e†ects, the data to constrain the properties of the dust and show that resulting IRAS Sky Survey Atlas (ISSA: Wheelock et al. no power-law emissivity model can consistently explain the 1994) images are contaminated by signiÐcant striping and full spectral range of the dust emission. However, we Ðnd poor control of large-scale gradients. excellent agreement with a two-component model whose The Di†use Infrared Background Experiment (DIRBE) components we tentatively refer to as silicate- and carbon- on the COBE satellite is the perfect complement to IRAS.It dominated grains. With this model for the dust emissivity has relatively low angular resolution(0¡.7) but superbly con- function, extrapolation of Galactic dust emission from 100 trolled zero points and gains. This has led to the generation km to lower frequencies is based upon the Ðltered DIRBE of a map of the far-infrared sky with unprecedented accu- 100/240 km color temperature. racy and uniformity of coverage. Schlegel, Finkbeiner, & In ° 2, we discuss the COBE data sets and the details of Davis (1998, hereafter SFD98) created a merged map of the comparisons using SFD98. Section 3 explores a variety of IRAS and DIRBE data with an angular resolution of 6@ and one-component dust models, demonstrating that a single DIRBE-quality calibration. Their full-sky map shows the power-law emissivity fails to explain the data, as does a pervasive extent of the infrared cirrus and has proved suc- broadened temperature distribution. Section 4 explores a cessful for estimation of extragalactic reddening. But family of two-component dust models, in which energy equally important will be the use of this type of data for balance and the temperature of the separate components estimation of Galactic foreground for the coming gener- are tightly coupledÈone of which achieves excellent agree- ation of CMBR experiments, including MAP and Planck ment with the FIRAS data. Section 5 discusses the robust- and a host of ground- and balloon-based projects. ness of this best model with respect to various ISM 867 868 FINKBEINER ET AL. Vol. 524 environments, and °6 compares our predictions to (DMR) of 123 frequency bins at 100 \l\2100 GHz (140 microwave observations, demonstrating that the micro- km \j\3 mm). Note that data in the lowest two fre- wave emission may exceed the predictions of any thermal quency bins are o† the page in some of the Ðgures but are (vibrational) emission mechanisms. This is perhaps the sig- used in the Ðts. nature of spinning dust grains emitting electric dipole radi- ation (Draine & Lazarian 1998b) or the signature of 2.2. DMR Data free-free emission. Summary and conclusions are presented DMR observed the sky at three frequencies, 31.5, 53, and in ° 7. Details of the frequency bin choice and recalibration 90 GHz, achieving the Ðrst detection of anisotropy in the can be found in Appendix A. Appendix B contains details CMBR (Smoot et al. 1992). In this paper we use the 4-Year on computational methods, and Appendix C provides DMR Skymaps dated 1995 April 18, which have the mono- details on data presentation. pole and dipole removed. These maps do not inÑuence any of our model Ðts, but are compared with our predictions in 2. DATA SETS ° 6. Kogut et al. (1996) observed a correlation between Galac- The COBE (COsmic Background Explorer) satellite con- tic dust and the 31.5 and 53 GHz channels of DMR that is sisted of three instruments, DMR (Di†erential Microwave much greater than that expected from any models of Radiometer), FIRAS (Far Infrared Absolute thermal (vibrational) emission by dust. Alternative explana- Spectrophotometer), and DIRBE (Di†use Infrared Back- tions such as spinning dust grains (Draine & Lazarian ground Experiment). In this paper we shall compare predic- 1998b) or spatially correlated free-free emission have been tions of dust emission based on DIRBE in the far-infrared proposed but are not well constrained by existing data (cf. with that observed by FIRAS at lower frequencies. In addi- de Oliveira-Costa et al. 1998). We discuss this excess emis- tion, we extend this correlation to still lower frequencies sion in ° 6. (31.5, 53, and 90 GHz) observed by DMR. Although the DMR Ñuctuations are dominated by intrinsic CMBR 2.3. DIRBE Data and SFD Dust Maps anisotropy, a residual correlation with DIRBE is detectable 2.3.1. SFD Emission Map even at high latitudes. SFD98 presented a full-sky 100 km cirrus emission map constructed from both the DIRBE and IRAS/ISSA data 2.1. FIRAS Spectra sets. The map is well calibrated, zodiacal lightÈsubtracted, The objective of the FIRAS instrument was to compare Fourier-destriped, and point sourceÈsubtracted, with a Ðnal the cosmic microwave background radiation (CMBR) to an resolution of6.1.@ Complete descriptions of these maps may accurate blackbody and to observe the dust and line emis- be found in SFD98.2 For comparisons with FIRAS and sion from the Galaxy. It is a polarizing Michelson interfer- DMR, the full resolution of the IRAS/DIRBE map is not ometer (Mather 1982), operated di†erentially with an required. Instead, we use the0¡.7 DIRBE map with zodiacal internal reference blackbody and calibrated by an external light removed as described in SFD98, with point sources blackbody with an emissivity known to better than one part included.