1: \documentclass[preprint2]{aastex}
2:
3: %\usepackage{url}
4: \usepackage{natbib}
5:
6: \slugcomment{To appear in the Astrophysical Journal}
7:
8: \shorttitle{BLAST05: Targeted Sources}
9: \shortauthors{Truch, M.~D.~P.~et al.}
10:
11: \newcommand{\h}{^{\mathrm h}}
12: \newcommand{\m}{^{\mathrm m}}
13: \newcommand{\s}{^{\mathrm s}}
14: \newcommand{\IRAS}{{\it IRAS}}
15: \newcommand{\ISO}{{\it ISO}}
16: \newcommand{\Spitzer}{{\it Spitzer}}
17: \newcommand{\Herschel}{{\it Herschel}}
18: \newcommand{\E}[1]{\ensuremath{\times 10^{#1}}}
19:
20: \begin{document}
21:
22: \title{The Balloon-borne Large Aperture Submillimeter Telescope (BLAST) 2005:
23: Calibration and Targeted Sources}
24:
25: \author{
26: M.~D.~P.~Truch,\altaffilmark{1}
27: P.~A.~R.~Ade,\altaffilmark{2}
28: J.~J.~Bock,\altaffilmark{3,4}
29: E.~L.~Chapin,\altaffilmark{5}
30: M.~J.~Devlin,\altaffilmark{6}
31: S.~Dicker,\altaffilmark{6}
32: M.~Griffin,\altaffilmark{2}
33: J.~O.~Gundersen,\altaffilmark{7}
34: M.~Halpern,\altaffilmark{5}
35: P.~C.~Hargrave,\altaffilmark{2}
36: D.~H.~Hughes,\altaffilmark{8}
37: J.~Klein,\altaffilmark{6}
38: G.~Marsden,\altaffilmark{5}
39: P.~G.~Martin,\altaffilmark{9,10}
40: P.~Mauskopf,\altaffilmark{2}
41: C.~B.~Netterfield,\altaffilmark{10,11}
42: L.~Olmi,\altaffilmark{12,13}
43: E.~Pascale,\altaffilmark{11}
44: G.~Patanchon,\altaffilmark{14}
45: M.~Rex,\altaffilmark{6}
46: D.~Scott,\altaffilmark{5}
47: C.~Semisch,\altaffilmark{6}
48: C.~Tucker,\altaffilmark{2}
49: G.~S.~Tucker,\altaffilmark{1}
50: M.~P.~Viero,\altaffilmark{10}
51: D.~V.~Wiebe\altaffilmark{11}
52: }
53:
54:
55: \altaffiltext{1}{Department of Physics, Brown University, 182 Hope Street, Providence, RI 02912;
56: {\url{matthew@truch.net}}}
57: \altaffiltext{2}{Department of Physics \& Astronomy, Cardiff University, 5 The Parade, Cardiff, CF24~3AA, UK}
58: \altaffiltext{3}{Jet Propulsion Laboratory, Pasadena, CA 91109-8099}
59: \altaffiltext{4}{Observational Cosmology, MS 59-33, California Institute of Technology, Pasadena, CA 91125}
60: \altaffiltext{5}{Department of Physics \& Astronomy, University of British
61: Columbia, 6224 Agricultural Road, Vancouver, BC V6T~1Z1,
62: Canada}
63: \altaffiltext{6}{Department of Physics and Astronomy, University of Pennsylvania, 209 South 33rd Street,
64: Philadelphia, PA 19104}
65: \altaffiltext{7}{Department of Physics, University of Miami, 1320 Campo Sano Drive, Coral Gables, FL 33146}
66: \altaffiltext{8}{Instituto Nacional de Astrof\'{\i}sica \'Optica y Electr\'onica (INAOE),
67: Aptdo. Postal 51 y 72000 Puebla, Mexico}
68: \altaffiltext{9}{Canadian Institute for Theoretical Astrophysics, University of Toronto,
69: 60 St. George Street, Toronto, ON M5S~3H8, Canada}
70: \altaffiltext{10}{Department of Astronomy \& Astrophysics, University of Toronto, 50 St. George Street,
71: Toronto, ON M5S~3H4, Canada}
72: \altaffiltext{11}{Department of Physics, University of Toronto, 60 St. George Street, Toronto, ON M5S~1A7, Canada}
73: \altaffiltext{12}{Istituto di Radioastronomia, Largo E. Fermi 5, I-50125, Firenze, Italy}
74: \altaffiltext{13}{University of Puerto Rico, Rio Piedras Campus, Physics Dept., Box 23343, UPR station,
75: San Juan, Puerto Rico}
76: \altaffiltext{14}{Laboratoire APC, 10, rue Alice Domon et L{\'e}onie Duquet 75205 Paris, France}
77:
78: \begin{abstract}
79: The Balloon-borne Large Aperture Submillimeter Telescope (BLAST) operated
80: successfully during a 100-hour flight from northern Sweden in June 2005
81: (BLAST05)\@. As part of the calibration and pointing procedures, several
82: compact sources were mapped, including solar system, Galactic, and
83: extragalactic targets, specifically Pallas, CRL~2688, LDN~1014,
84: IRAS~20126+4104, IRAS~21078+5211, IRAS~21307+5049, IRAS~22134+5834,
85: IRAS~23011+6126, K3-50, W~75N, and Mrk~231. One additional source, Arp~220,
86: was observed and used as our primary calibrator. Details of the overall BLAST05
87: calibration procedure are discussed here. The BLAST observations of each
88: compact source are described, flux densities and spectral energy distributions
89: are reported, and these are compared with previous measurements at other
90: wavelengths. The 250, 350, and 500\,\micron\ BLAST data can provide useful
91: constraints to the amplitude and slope of the submillimeter continuum, which in
92: turn may be useful for the improved calibration of other submillimeter
93: instruments.
94: \end{abstract}
95:
96: \keywords{balloons --- submillimeter --- telescopes}
97:
98: \section{Introduction} \label{sec:intro}
99: The June 2005 flight of the Balloon-borne Large Aperture Submillimeter
100: Telescope (BLAST) incorporated a 2-m spherical primary mirror and large-format
101: bolometer arrays operating at 250, 350, and 500\,\micron. A complete
102: description of the BLAST instrument is given in \citet{pascale2008}. The BLAST
103: wavelengths sample the peak of the spectral energy distribution (SED) for cool
104: dust (${\sim}\,10$--40\,K) and are in a regime which is difficult or impossible
105: to access from even the best ground-based sites. As a result, BLAST has the
106: ability to conduct unique Galactic and extragalactic submillimeter surveys with
107: arcminute resolution and high sensitivity. BLAST's primary scientific
108: motivations are to study the angular and redshift distribution and evolution of
109: high-redshift star-forming galaxies and to identify proto-stellar cores and the
110: earliest stages of star formation within Galactic molecular clouds.
111:
112: BLAST conducted a 100-hour flight, launching from northern Sweden on 2005 June
113: 12, and landing in northern Canada on 2005 June 16 (BLAST05)\@. One
114: relatively shallow
115: extragalactic field and several large (${\sim}\,8\,{\rm deg}^2$) Galactic fields
116: \citep[see][]{chapin2008, hargrave2008} were mapped, from which a large number
117: of Galactic pre-stellar sources were extracted at very high signal-to-noise
118: ratio. The results obtained from the BLAST flight from Antarctica in December
119: 2006 (BLAST06) will be discussed in future articles.
120:
121: BLAST05 continuum observations were made with three focal plane arrays
122: consisting of 139, 88, and 43 detectors at 250, 350, and 500\,\micron,
123: respectively. The optics and detector layout in the cryostat make
124: simultaneous measurements possible by three arrays, having a common field of view of
125: about 14\arcmin\ $\times$ 7\arcmin. The detectors are silicon-nitride
126: micromesh (``spider-web'') bolometric detectors coupled with $2f\lambda$
127: \citep[maximum optical efficiency,][]{griffin2002}
128: spaced feedhorn arrays \citep{turner2001}. The arrays are prototypes of those
129: developed for the SPIRE instrument on \Herschel\ \citep{griffin2004}.
130:
131: In this paper we report the observations of compact sources that were targeted
132: by BLAST largely for the purposes of flux and pointing calibration. These
133: sources include Pallas, CRL~2688, K3-50, W~75N, Mrk~231, and Arp~220, which have
134: all been well studied in the submillimeter and are commonly used as primary or
135: secondary standards. We also conducted several observations of known bright
136: proto-stars that were not otherwise included in our wider Galactic plane survey
137: fields, specifically IRAS~20126+4104, IRAS~21078+5211, IRAS~21307+5049,
138: IRAS~22134+5834, and IRAS~23011+6126. Finally we observed the ``starless''
139: dense core LDN~1014 that was recently studied with the {\it Spitzer Space
140: Telescope}\ \citep{young2004}.
141:
142: Together these observations represent a broad sampling of bright submillimeter
143: sources that were available during the flight, given the strict visibility
144: constraints of the telescope \citep{pascale2008}. The planets Mars and Uranus,
145: often used as submillimeter calibrators, were not visible during this flight.
146: Although observations of Saturn were attempted, there were concerns about gain
147: stability because it was relatively close to the direction of the Sun, and it may
148: also have suffered from saturation effects in the detectors, and hence we
149: abandoned attempts to use these particular observations. After carefully
150: assessing the systematic effects in the relevant parts of the SEDs of the other
151: sources, we chose to single out Arp~220 as our primary flux calibrator.
152: Using Arp~220, we determine that the BLAST05 calibration uncertainties are 12\%,
153: 10\%, and 8\% in the 250, 350, and 500\,\micron\ bands, respectively.
154:
155: This paper is divided into three main sections. Section~\ref{sec:red}
156: outlines the basic reduction steps and characterization of BLAST05
157: data. Section~\ref{sec:perf} discusses the performance of the warm
158: optics in BLAST05 and the impact of the degraded performance on the
159: resolution and sensitivity of the experiment.
160: Section~\ref{sec:fluxcal} describes in detail the absolute calibration derived from
161: the primary flux-calibrator Arp~220, while Section~\ref{sec:obs}
162: summarises the BLAST flux-densities and the simple fits of a
163: modified black-body emission model to the SEDs of each of
164: the targeted sources from the BLAST05 flight.
165:
166: \section{Data Reduction} \label{sec:red}
167: We now describe the main steps in the data reduction and calibration
168: process. Raw data from BLAST consist of a set of bolometer timestreams in
169: voltage units, sampled at $100\,$Hz.
170: The raw bolometer data are first cleaned for post-flight analysis. The
171: data are de-spiked and then deconvolved to remove the electronics filters from
172: the timestreams; see \citet{pantachon2008} for specifics of the cleaning and
173: deconvolution performed on the BLAST data. The cleaned data are combined with
174: a post-flight pointing solution \citep{pascale2008} to make maps at each
175: frequency, taking advantage of the multiple detectors, as well as significant
176: scan cross-linking, to minimize striping due to instrumental drifts.
177:
178: The bolometers in each array are corrected for relative gains, or flat-fielded, so that meaningful
179: multi-bolometer maps can be generated. The flat-field corrections are
180: determined using individual maps made for each bolometer from a single
181: point-source calibrator, in this case, CRL~2688 (the brightest point-like source observed).
182: These scans were designed
183: such that a fully-sampled map can be generated from each bolometer
184: individually. For each bolometer, the total flux from the point source is
185: integrated and a flat-fielding coefficient for each bolometer is then calculated
186: as the ratio of the flux from that bolometer and the flux from an arbitrarily
187: chosen reference bolometer. All subsequent maps are generated from timestreams
188: that apply the flat-fielding coefficient appropriately to each bolometer.
189:
190: To calculate the flux density from a point source, we adopted a
191: matched-filtering technique similar to that used to extract point sources from
192: several recent extra-galactic submillimeter surveys
193: \citep[e.g.,][]{coppin2006,scott2006}. A point spread function (PSF) is
194: generated by stacking and averaging several point sources from various maps in
195: telescope coordinates (azimuth and elevation relative to the telescope).
196: We area-normalize the PSF, $P,$ such that
197: \begin{equation}
198: \sum_{x,y} P(x,y)\delta x \delta y = 1,
199: \end{equation}
200: where the double sum is over all pixels in the PSF map, and $\delta x$ and
201: $\delta y$ are the angular dimensions of a pixel such that the units of $P$ are
202: sr$^{-1}$. In all BLAST maps, square pixels are used so that
203: $\delta x = \delta y$. BLAST maps are calibrated in surface brightness units
204: [Jy sr$^{-1}$]. A map of a point source, $M(x,y)$, can be modeled by the
205: normalized PSF centered over the source, scaled by its flux density, $S$, or
206: $S P(x,y) \simeq M(x,y)$. If each pixel in $M(x,y)$ has an uncertainty
207: $\sigma(x,y)$, calculated in the map-making process \citep{pantachon2008},
208: then we can write $\chi^2$ for the model, assuming independant noise:
209: \begin{equation}
210: \chi^2 = \sum_{x,y}\frac{\left(S P(x,y) - M(x,y)\right)^2}{\sigma^2(x,y)}.
211: \end{equation}
212: Minimizing this $\chi^2$ results in the maximum likelihood flux density,
213: \begin{equation} \label{eq:flux}
214: S = \frac{\sum_{x,y}(M(x,y)P(x,y)/\sigma^2(x,y))}
215: {\sum_{x,y}(P(x,y)/\sigma(x,y))^2}.
216: \end{equation}
217:
218: This technique uses our knowledge of the beam as a model for the
219: shape of un-resolved point-sources in BLAST maps. Fitting the amplitude
220: of this template is optimal in a S/N sense, and particularly important
221: for measurements of faint sources. This method produces smaller
222: measurement uncertainties than simple aperture photometry, as pixels
223: near the peak signal are weighted more heavily than pixels in the
224: wings of the brightness distribution (aperture photometry weights all
225: pixels equally).
226:
227: This technique gives a low statistical uncertainty both because we have
228: intrinsically high signal-to-noise data and also because knowledge of the beam
229: shape is used to weight the uncertainties; however, errors in the PSF will bias
230: the result. To check for a bias we used simple aperture photometry as a second
231: measure of flux densities for CRL~2688. The difference between the matched
232: filter and aperture photometry flux densities of CRL~2688 is less than 3\%.
233: Since the matched filter provides significantly smaller statistical errors, it
234: is used to extract flux densities from all point-like sources in this paper.
235:
236: The PSF for each individual bolometer varies slightly across the
237: array. The stacked PSF takes this effect into account for the bulk of
238: the map over which each bolometer contributes approximately equally to
239: the combined signal. However, the extreme edges of the map are only
240: sampled by a fraction of the bolometers and therefore exhibit
241: different effective PSFs. We extract flux densities from maps made
242: with single bolometers using the stacked PSF to assess the error
243: introduced by adopting this incorrect template, and find at most a
244: 10\% bias. The bias is maximized at the very edges of the map, but
245: affects flux densities measured to within 14\arcmin\ (width of the
246: array) of the edge of the map along the scan direction (approximately
247: aligned with R.A.), and 7\arcmin\ (height of the array) in the
248: transverse direction (approximately aligned with Dec.). None of the
249: measurements discussed in this paper are affected by this bias.
250:
251: The flat-fielding process was repeated for CRL~2688 observations made during the
252: middle and end of the flight to check for stability. Although no bias or trend
253: is seen, the coefficients for each individual bolometer vary by 5\% rms.
254: This effect averages down when multiple
255: bolometers are used, and is negligible in full bolometer-array maps as used in
256: this paper. The 5\% scatter in flatfield coefficients appears to be dominated
257: by measurement uncertainties; therefore this value can be considered a very
258: robust upper-limit on the systematic drift in the relative calibration, and is
259: insignificant compared to the total calibration error budget given in
260: Table~\ref{calib}.
261:
262: \subsection{Responsivity Variations}
263: Variations in bolometer loading (due to changing sky emission, for example) and
264: bolometer base-plate temperature cause changes in bolometer responsivities. To
265: characterize and correct for these changes, a calibration lamp (``cal-lamp'') is
266: located in the optics box at the center of the reflective Lyot stop \citep[see
267: \S 3 of][]{pascale2008}, allowing for the measurement of these changes. The
268: cal-lamp is pulsed once every 10--15 minutes during flight. The cal-lamp is of
269: the same design as the one used for SPIRE \citep{hargrave2006}. It is designed
270: to provide a stable signal over the flight, so any change in cal-lamp amplitude
271: as measured by a bolometer is directly proportional to a change in that
272: bolometer's responsivity. The signal baseline is removed by fitting a line to
273: a 550\,ms segment of data before and after the cal-lamp pulse. A template
274: cal-lamp profile from a raw bolometer timestream is fit to every cal-lamp pulse
275: in every bolometer. The template is chosen from a typical bolometer and
276: it has been verified that the template fit amplitude is within 2\% of a
277: square-wave fit. The amplitude of the fit is interpolated
278: over time and inverted to generate a multiplicative scaling cal-lamp timestream
279: for each bolometer. Fluctuations in the responsivity are less than 8\% rms,
280: dominated by diurnal variations due to differences in atmospheric loading and
281: thermal emission from the telescope's changing temperatures. On time scales of
282: a typical source map the fluctuations are less than 2\% rms. A constant
283: responsivity timestream for each bolometer is generated by applying the
284: multiplicative scaling cal-lamp timestream. The large-scale fluctuations are
285: suppressed and the final responsivity variations are less than 2\% rms.
286:
287:
288: \section{BLAST05 Warm Optics Performance} \label{sec:perf} During the
289: BLAST05 flight, the warm optics (primary and secondary mirrors) did not perform
290: within the specifications. The beams as designed were expected to be close to
291: diffraction limited and approximately Gaussian,
292: $\exp(-\theta^2/2\sigma_{\rm B}^2)$,
293: with $\rm{FWHM} = 2 (2 \ln 2)^{1/2} \sigma_B =$ 32\arcsec,
294: 48\arcsec, and 64\arcsec, at 250, 350, and 500\,\micron, respectively. In this
295: case the full width at half power of the beam, FWHP, is equal to the FWHM, and
296: the beam solid angle can be characterized as
297: $\Omega = \pi \alpha_{\rm D}^2/(4 \ln 2)$ with $\alpha_{\rm D} = \rm{FWHP}$.
298: The point-spread functions in all three BLAST bands were measured from
299: multiple observations of the proto-planetary nebula CRL~2688.
300: The in-flight beam-shapes are shown in Figure~\ref{psfs}. Noting that the
301: central structure in each is on a scale comparable to the diffraction limit,
302: the beams are clearly far from ideal, distributing considerable power into an
303: outer ring (hexagon) of diameter ${\sim}\,200$\arcsec.
304:
305: \begin{figure*}[t]
306: \includegraphics[angle=0, width=6.5in]{f1.eps}
307: \caption{ Point Spread Functions (PSFs), provided by observations of
308: CRL~2688 (\S~\ref{crl}), for each of the three wavebands from BLAST05,
309: generated by stacking several point source maps in telescope
310: coordinates. The small circles represent the expected diffraction
311: limited FWHM for each of the wavebands (Pascale et al. 2008). The
312: PSFs are significantly more extended than expected, reducing
313: sensitivity to point sources and the ability to distinguish crowded
314: sources amidst fluctuating cirrus.
315: \label{psfs}
316: }
317: \end{figure*}
318:
319: With the pre-flight predicted Noise Equivalent Flux Density (NEFD) of 200
320: mJy\,${\mathrm s}^{1/2}$ we would expect 1-$\sigma$ surface brightness
321: fluctuations at the nominal resolutions of
322: ${\sim}\,\rm{NEFD}/\Omega $
323: or 11, 4.7, and 2.6 MJy sr$^{-1}\,{\mathrm s}^{1/2}$ at 250, 350, and
324: 500\,\micron, respectively.
325: The measured sensitivities were in fact 8.8, 4.8, and
326: $2.7\,{\rm MJy\,sr^{-1}}\,{\mathrm s}^{1/2}$, respectively,
327: which shows that BLAST did not suffer a reduced
328: sensitivity to extended diffuse sources, i.e. detectors worked as planned and optical
329: efficiency was as planned.
330:
331: There is, however, a marked loss in sensitivity for point sources, because of
332: the larger-than-expected beam solid angle. The FWHP of the degraded beam is
333: ${\sim}\,200$\arcsec\ (somewhat smaller at $500\,\mu$m), given the power in the
334: ring. The reduction in point-source sensitivity is
335: ${\lesssim}\,(200/32)^2\,{\simeq}\,40$ at
336: 250\,\micron. This is a pessimistic estimate because there is
337: considerable higher-resolution structure in the beam, which one takes advantage of in the
338: PSF fitting described above. Alternatively, one can examine
339: the histogram of pixel brightnesses contributing to the PSF, and
340: working from the brightest, find the beam area that accounts for half the power.
341: We find $\alpha_{\rm D}
342: = $ 186\arcsec, 189\arcsec, and 189\arcsec\ for 250, 350, and 500\,\micron,
343: respectively, and so the sensitivity is reduced by factors of
344: approximately 34, 16, and 9.
345: Despite this reduction, the targeted
346: sources described in this paper were still detected with high signal to noise.
347: Further data from BLAST05 are discussed in \citet{chapin2008},
348: \citet{hargrave2008}, and other forthcoming papers.
349:
350: Although it is not critical to the analysis here, we have investigated the
351: origin of the beam degradation. One possibility is misalignment,
352: shifting, and tilting of the secondary and primary with respect to the nominal
353: optical axis. The launch was sufficiently violent to cause
354: the inner-frame locking mechanism to fail; the inner-frame (which houses
355: the primary, secondary, and camera) violently hit both hardware
356: elevation stops, which could have damaged the telescope, including the
357: secondary support mechanism and the primary itself.
358: %
359: From visual inspection of the images (and power spectra) there is clearly
360: information in the beams near the diffraction limit, such as the central dip at
361: 350\,\micron, which is not present at the other two wavelengths, and the width
362: of the outer-ring. While these structures are suggestive a problem in the telescope,
363: numerous experiments with the
364: optical design software ZEMAX\footnote{\url{http://www.zemax.com/}}, in which
365: all the optical elements were shifted and/or tilted from their nominal
366: position, we were unable to reproduce the distinctive features in the beams
367: at all the wavelengths simultaneously.
368: %
369: There appears to have been an optical problem beyond out-of-focus or
370: out-of-alignment optics.
371: %
372: The hexagonal shape of the beam does suggest damage to the
373: carbon-fiber primary itself, which was constructed from six panels
374: (segments). In addition to effects of the violent launch, another contributing
375: factor could have been the light rain the payload endured before
376: launch. Carbon fiber becomes particularly weak in a high humidity
377: environment, possibly leading to delamination of the six panels either
378: before or during the launch. Furthermore, there was the usual
379: freeze-thaw cycle on ascent, where the ambient temperature dropped as
380: low as $-65$\,\degr{}C.
381: %
382: Further analysis of the PSF (including Fourier Transform) was not performed as
383: phase-information is required for proper analysis; as the mirror was fully
384: destroyed upon landing there is little use in debugging the system further.
385:
386:
387: \section{Astronomical Flux-calibration} \label{sec:fluxcal}
388:
389: The primary scientific goals of the BLAST experiment demanded a flux
390: calibration-accuracy of better than 10\% in all three BLAST
391: pass-bands. Achieving this was complicated by the variable
392: sky-visibility of BLAST due to the unstable projected-latitude of the
393: telescope gondola during the flight, and the restrictions on
394: visibility due to the Sun and Moon avoidance criteria, the orientation
395: of sun-shields and other baffling, and the elevation range
396: (25--60\degr{}) of the gondola's inner-frame
397: \citep{pascale2008}. Consequently BLAST had only limited access to the
398: calibration sources commonly used at submillimeter and far infrared (FIR)
399: wavelengths.
400:
401: Since the ecliptic-plane was not visible during the BLAST05 flight, no
402: absolute flux-calibration of BLAST could be determined from
403: observations of Uranus or Mars, for which model SEDs are known to have
404: systematic uncertainties ${<}\,5$\% at submillimeter wavelengths
405: \citep{griffin1993,wright2007}. The pre-flight strategy for achieving
406: a 10\% calibration accuracy, recognising the above visibility
407: constraints, therefore forced us to identify other Galactic and
408: extragalactic sources that could act as primary and secondary
409: calibrators, with the following requirements: (i) availability
410: throughout the flight; (ii) already considered, in some cases, as
411: secondary calibrators for ground-based sub-mm telescopes and FIR
412: satellites; (iii) well-constrained SEDs in the FIR to mm-wavelength
413: regime, enabling accurate interpolation of the band-averaged
414: flux-densities at BLAST wavelengths; (iv) bright ($\gg 1$~Jy at
415: 500\,\micron) and compact sources (with respect to the BLAST beam-size,
416: $< 20$\arcsec) that reside in regions with minimal spatial-structure
417: in the Galactic foregrounds or backgrounds, allowing accurate
418: subtraction of any extended emission. Given these criteria we
419: scheduled regular observations throughout the flight of bright
420: embedded protostellar-sources and compact \ion{H}{2} regions within the W58
421: and Cygnus-X molecular-cloud complexes (including K3-50, W~75N, and
422: DR21), a post-AGB star with planetary nebula (CRL~2688), the asteroid
423: 2~Pallas, and two ultraluminous infra-red galaxies (ULIRGs), Arp~220 and Mrk~231.
424:
425: In the following sub-section we justify our selection of the ULIRG
426: Arp~220 as our primary calibration source for the BLAST05 flight.
427:
428:
429: \begin{figure}[t]
430: \begin{center}
431: \includegraphics[width=3in]{f2.eps}
432: \caption{ Spectral Energy Distribution (SED) of Arp~220, the absolute
433: flux calibrator for BLAST05. The best-fit model (heavy solid-line) was
434: constrained by the published data (blue triangles, discussed in
435: Section~\ref{sec:arp220}), and excluded BLAST05 measurements. The
436: grey-lines shows the 68\% confidence interval, estimated from 100
437: Monte-Carlo simulations, about the best-fit model. Black diamonds
438: indicate the model predictions for BLAST05 at 250, 350 and
439: 500\,\micron. The uncertainties associated with these predictions are
440: 11\%, 10\%, and 8\% in each band, respectively. The best-fit
441: parameters of the single-temperature dust model are given in
442: Table~\ref{fits}.
443: \label{sed:arp220}
444: }
445: \end{center}
446: \end{figure}
447:
448:
449: \subsection{Arp~220 - a primary flux-calibrator for BLAST05} \label{sec:arp220}
450: Arp~220 \citep{arp1966} is a well-studied
451: ULIRG at redshift $z=0.018$ with a FIR
452: luminosity of approximately $2 \times 10^{12} $\,L$_{\odot}$. The
453: sub-millimeter emission is believed to be due to dust heated primarily
454: by active star-formation (at a rate of
455: $\sim 200 {\mathrm M}_{\odot}/{\mathrm yr}$) and to
456: a lesser extent by an embedded Active Galactic Nucleus (AGN). This
457: emission is confined to a compact ($< 2$\arcsec\ diameter) region, as
458: traced through interferometric imaging, by the distribution of the
459: 1.3\,mm dust continnum, and the molecular gas in the CO(2--1)--line
460: \citep{scoville1997} that fuels the starburst activity.
461: Arp~220 resides far from the Galactic plane ($b = +53\degr$) and
462: extended foreground Galactic contamination as seen in the 100 \micron\
463: map \citep{schlegel1998} does not affect the measured SED\@.
464: Arp~220 is therefore a point-like source as observed by BLAST, as well as
465: most ground-based and space-borne FIR to millimeter telescopes, making
466: the construction of the FIR-to-millimeter SED insensitive to the
467: aperture-sizes of the photometric measurements
468: ({Figure~\ref{sed:arp220}).
469:
470: The quoted measurement errors for the ground-based 350--1100\,\micron\
471: photometry of Arp~220 are ${\leq}\,$\,10\%, while the absolute
472: calibration uncertainties are typically 10--20\%. In the
473: case of shorter-wavelength FIR measurements from ISOPHOT at
474: 60--200\,\micron\ \citep{klaas2001} and 120--200\,\micron\
475: \citep{spinoglio2002}, the absolute calibration of ISOPHOT, which is
476: estimated to be ${\sim}\,$30\%, also dominates the total photometric
477: accuracy. More precise photometry, however, at 60 and 100\,\micron\
478: is available from the \IRAS\ Revised Bright Galaxy Sample
479: \citep[RBGS,][]{sanders2003}\@, with statistical uncertainties that
480: have been quoted at the $\sim$\,0.1\% level.
481: The difference between the absolute calibration of the RBGS and the
482: original IRAS Bright Galaxy Samples ranges from 5--25\%, and our
483: analysis therefore assumes an additional 10\% absolute calibration
484: uncertainty on the \IRAS\ flux densities. As we show below, combining
485: all the above data, and accounting for systematic uncertainties, we
486: have been able to constrain a simple SED model of thermal dust
487: emission that allows interpolated band-averaged flux-densities of
488: Arp~220 to be estimated across the BLAST bandpasses with an acceptable
489: $\sim$\,10\% uncertainty (Figure~\ref{sed:arp220}).
490:
491: In order to predict the BLAST band-averaged fluxes of Arp~220, a
492: single-temperature modified-blackbody SED of the form $S_\nu = A
493: (\nu/\nu_0)^\beta B_\nu(T)$ was first fit to existing submillimeter
494: and FIR observations described previously. In the above expression,
495: $S_\nu$ is the flux density, $A$ is the amplitude, $\beta$ is the
496: emissivity-index of the radiating dust-grains, $\nu_0$
497: is fixed at $c / 250$\,\micron,
498: and $B_\nu(T)$ is the Planck function for a
499: blackbody with temperature $T$.
500:
501:
502: The smooth SED is fit to the data using $\chi^2$ minimization.
503: In addition to the measurement errors quoted in \S
504: \ref{sec:arp220}, we have assumed an additional {\it correlated}
505: error of 5\% for all of the photometry, since the bulk of these instruments
506: were calibrated using the same Uranus SED, which is known to have an
507: uncertainty of $\sim$\,5\% \citep{griffin1993}.
508: Furthermore, we have been careful to account for correlated errors using
509: the full data covariance matrix,
510: \begin{equation}
511: \chi^2 = (\tilde{{\mathbf S}} - {\mathbf S}) C^{-1}
512: (\tilde{{\mathbf S}} - {\mathbf S})^T,
513: \label{eq:chisquared}
514: \end{equation}
515: where ${\mathbf S}$ is the vector of measured flux densities, and
516: $\tilde{{\mathbf S}}$ are the model predictions. The diagonal
517: elements of $C$ give the variances, $\sigma_i^2$, calculated as the
518: quadrature sums of the photometric and calibration uncertainties
519: described above. The off-diagonal elements of $C$ are calculated as
520: follows. For each pair of measurements taken with the same group and
521: same instrument, $i$ and $j$,
522: calibration uncertainties are assumed to be completely correlated such
523: that $C_{ij} = \sigma_i^\mathrm{c} \sigma_j^\mathrm{c}$, where
524: $\sigma^\mathrm{c}$ is the standard deviation of the calibration error
525: component for the measurement. Monte Carlo simulations are performed
526: to characterize the range of models that are consistent with the data.
527: One thousand mock data sets are generated by adding realizations of
528: correlated Gaussian noise described by $C$ to the measured flux
529: densities ${\mathbf S}$. Each one of these data sets is then re-fit
530: using Equation~\ref{eq:chisquared}. The best-fit model ($T=41.7$\,K
531: and $\beta = 1.3$,
532: for the observed SED, and
533: its uncertainties, are shown in Figure~\ref{sed:arp220}.
534:
535: %
536: The band-averaged flux densities for each BLAST wavelength are obtained for
537: each SED model fit as
538: \begin{equation}
539: {\tilde S_{\mathrm B}} = \int T(\nu) {\tilde S}(\nu) d\nu,
540: \end{equation}
541: where we use tildes to represent model quantities. Here the normalized filter
542: transmission profiles $T(\nu)$ are measured empirically using a Fourier
543: transform spectrometer \citep{pascale2008}. From these 1000 sets of
544: band-averaged fluxes, we obtain estimates of variances of the calibrator
545: brightnesses in each band, $\left<\sigma^2_i\right>$, as well as the correlated
546: error between the fluxes in each band, $\left<\sigma_i \sigma_j\right>$.
547:
548: Maximum-likelihood estimates for the flux densities of all BLAST05
549: sources are obtained from equation~(\ref{eq:flux}) using un-calibrated
550: maps of Arp~220 in units of Volts. Multiple observations of CRL~2688
551: were made throughout the flight, allowing us to estimate systematic
552: gain variations not traced by the internal cal-lamp pulses at the 3\%
553: level. Since this gain variation is significantly lower than the
554: calibrator flux-uncertainty estimated above, we calculate single
555: calibration coefficients $c_i$ in each band for the entire flight as
556: the ratio of the band-averaged fluxes to the raw fluxes from our maps.
557: Note that this calibration depends explicitly on the assumption of a
558: smooth thermal SED that neglects the effects of molecular emission
559: lines.
560:
561: This problem has been examined in detail for SCUBA 850\,\micron\ continuum
562: observations of a variety of Galactic objects. \citet{friesen2005}
563: conclude that the line contamination is typically $<10$\% in hot
564: Galactic molecular cores, and a similar conclusion was reached by
565: \citet{johnstone2003} in a study of protostellar sources in Orion,
566: although in some exceptional cases the line-contamination of the
567: 850\,\micron\ continuum fluxes in the most energetic sources reached
568: levels $>50$\%. Similarly, line-contamination, due exclusively to
569: CO(3--2) emission, has been measured to be $<$10\% in 850\,\micron\
570: continuum fluxes, for $\sim$\,60\% of the 45 local galaxies observed by
571: \citet{Yao2003}. Unfortunately there is almost no information on the
572: strengths of possible line-contamination in the BLAST pass-bands,
573: including redshifted molecular-lines at $\ll 300$\,\micron, and thus we
574: assume that $<10$\% line-contamination is present in the BLAST
575: data.
576:
577: Including all of these sources of uncertainty, the final percentage
578: errors (8--12\%) in the absolute flux calibration, and a Pearson
579: correlation matrix showing the relationship between errors in the
580: different bands, are given in Table~\ref{calib}. The calibration
581: coefficients are used to convert raw (Voltage) maps into calibrated
582: (Jy) maps. All astronomical analysis is performed on the calibrated
583: maps.
584:
585: \section{BLAST05 observations of bright sources} \label{sec:obs}
586:
587: In this section we summarise the calibrated BLAST observations of all
588: the bright targets presented in this paper, paying attention to the
589: level of agreement between their calibrated BLAST fluxes and the
590: predictions from an interpolation of their measured SEDs. Although
591: these bright-sources
592: were selected to provide checks on the consistency of the calibration
593: throughout the flight, and were not considered to be part of the
594: scientific program, we derive some of the physical properties
595: (temperatures, FIR luminosities, dust masses) that can be obtained
596: from a simple
597: model of thermal dust-emission that has
598: been fit to the improved spectral coverage of their SEDs (\S~\ref{sec:onetemp})
599:
600: In addition to conducting large-area surveys of Galactic fields
601: undergoing active starformation \citep{chapin2008},
602: BLAST05 observed a total of 9 bright, individual point-like sources
603: (Pallas, CRL~2688, IRAS~20126+4104, IRAS~21078+5211, IRAS~21307+5049,
604: IRAS~22134+5834, IRAS~23011+6126, Mrk~231, and Arp~220), complemented
605: by observations of more extended star-forming structures towards W75~N
606: in Cyg~X and K3-50 in W~58, as part of the overall calibration
607: strategy. Although not used explicitly as a calibration source, we
608: also present in this paper the BLAST05 observations of the compact low-mass
609: protostar L1014-IRS towards the nearby Galactic dark-cloud LDN~1014.
610:
611: All BLAST05 calibration targets are in the northern hemipshere
612: (23\degr $<$ Dec $<$ 61\degr; Table~\ref{fluxen}). BLAST therefore
613: provides valuable new photometric data at intermediate wavelengths to
614: complement those obtained in the FIR and the (sub-)millimeter regime.
615:
616: With the advantage of simultaneous observations at 250, 350, and
617: 500\,\micron, BLAST measures the thermal emission of these Galactic and
618: extragalactic objects with a high flux-calibration accuracy in a
619: wavelength-regime that traces the spectral curvature at wavelengths
620: longer than the spectral peak, and shorter than the Rayleigh-Jeans
621: limit for cold dust ($<50$\,K). In terms of modeling the SEDs, this
622: short-submillimeter wavelength-regime sampled by BLAST is critical for
623: unlocking the degeneracies between the effect of dust-temperature,
624: opacity and emissivity-index on the overall FIR-to-millimeter
625: emission. The SPIRE instrument, using similar bolometer-arrays and
626: identical filters to BLAST, following the future launch of {\it Herschel}
627: \citep{griffin2004}, will extend this work with greater sensitivity
628: and spatial resolution.
629:
630: It is not the purpose of this paper to provide a detailed account of physical
631: properties and nature of each bright-source observed in BLAST05. Thus we only
632: discuss briefly in the following sub-sections the collective results btained
633: from these BLAST observations at 250, 350, and 500\,\micron.
634:
635: \subsection{Spectral Energy Distributions}
636:
637: Flux densities at 250, 350, and 500\,\micron for all the above
638: point-sources are extracted using the matched-filter technique
639: (eq.~[\ref{eq:flux}] outlined in \S~\ref{sec:red}), while for
640: extended or blended-sources at the resolution of BLAST05 (e.g. towards
641: K3-50 and W~75N), we convolve the PSF with a simple model which is then fit
642: to the data. In the latter case the model consists of multiple
643: Gaussian sources, where we allow the amplitude, width, and position of
644: the Gaussians to vary as parameters of the fit.
645:
646: Since the BLAST filters have large spectral widths, and colors sampled
647: by the filters are a strong function of temperature (especially for
648: $T\la25$\,K), a correction must be made to calculate monochromatic
649: flux densities. Either a color correction may be applied to the
650: band-averaged flux densities, or effective filter wavelengths for each
651: object may be quoted. We choose the former, and quote effective flux
652: densities at precisely 250, 350, and 500\,\micron . Once the SED
653: has been fit to the BLAST data by minimizing $\chi^2$, to obtain the
654: best-fit dust temperature and emissivity-index, this is used to
655: calculate the correction
656: %
657: \begin{equation}
658: S(\nu) = S_{\mathrm{B}} \frac{\tilde{S}(\nu)}{\tilde{S}_{\mathrm{B}}},
659: \end{equation}
660: %
661: where $S_{\mathrm{B}}$ is the band-averaged BLAST flux
662: measurement, $\tilde{S}(\nu)$ is the
663: SED model flux density (evaluated at 250, 350, and 500\,\micron), and
664: $\tilde{S}_{\mathrm{B}}$ is the SED model band-averaged flux. The
665: BLAST color-corrected flux-densities for all bright sources presented in
666: this paper are given in
667: Table~\ref{fluxen}.
668:
669:
670: \subsubsection{Single-temperature cold dust models}\label{sec:onetemp}
671: Incorporating the BLAST measurements into the existing FIR--mm
672: wavelength ($\sim$ 50--2000\,\micron) SEDs for each source, we
673: fit only a single-temperature dust component while recognising that
674: any model that utilizes shorter-wavelength mid-IR data must
675: naturally include thermal emission from hot dust ($>$\,100\,K). This
676: decision is justified in the context of deriving the band-averaged
677: BLAST fluxes, which require only a smooth and accurate fit to the SED
678: over the restricted wavelength-range of the BLAST filters
679: \citep{pascale2008}. Although alternative and more complex models
680: (e.g. two-temperature components) can be fit to the same FIR--mm
681: data, the difference in the quality (e.g. minimised $\chi^2$) of
682: the fit is not significant
683: when compared to the single-temperature model.
684:
685: As previous studies of the FIR-mm SEDS of local galaxies
686: and ULIRGs have shown (e.g. Eales et al. 2000, Lisenfeld, Isaak \&
687: Hills 2000), relying on limited photometric data, with typically 3 to
688: 5 measurements in the FIR (from {\it IRAS} and {\it ISO}) and
689: (sub-)millimeter data (at
690: 850\,\micron\ and 1.2\,mm data), these SEDs can be well-represented by a
691: single dust temperature in the range of 25--50\,K. Similar conclusions have
692: been reached from the studies of dust in the Galactic ISM\@.
693: Even with the addition of ground-based 350 and 450\,\micron\ data
694: \citep[e.g, ][]{benford1999, hunter2000, dunne2001, beelen2006, coppin2008}
695: single temperature fits work fairly well, and there is typically no need to
696: fit multiple temperature components to SEDs unless one also brings in data at
697: $\lambda \ll$ 100\,\micron. More complicated fits would only be demanded with
698: higher-fidelity spectral measurements covering a much wider range of wavelengths.
699:
700: Given this, we show the mid-IR to millimeter-wavelength SEDs of all
701: the bright compact sources observed by BLAST05 and their best-fit
702: single-temperature models in
703: Figs.~\ref{sed:arp220}--\ref{sed:protostellar}, and present the
704: best-fit model-parameters in Table~\ref{fits}.
705:
706: Adopting a distance to each source (\S~\ref{sec:sources}), we determine the
707: FIR luminosity and dust mass from the SED fits, recognising that the
708: adoption of two or more temperature components would modify the
709: dust-mass estimate.
710: %
711: Determination of bolometric fluxes and dust masses follows the procedures
712: described in \citet{chapin2008}.
713: Uncertainties in quantities derived from the SED fits in
714: Table~\ref{fits} are obtained from Monte Carlos similar to those
715: described in \S~\ref{sec:arp220}, which now include BLAST data and their
716: correlated calibration uncertainties. For each mock data set the
717: quantity in question is derived from the model fit. These values are
718: then placed in histograms in order to extract means and 68\%
719: confidence intervals. Further details on these Monte Carlo
720: simulations are provided by \citet{chapin2008}.
721:
722:
723: \begin{figure}[t]
724: \begin{center}
725: \includegraphics[width=3in]{f3.eps}
726: \caption{CRL~2688 spectral energy distribution, showing BLAST data and
727: best-fitting models. Symbols and lines are as those described
728: in Fig.~\ref{sed:arp220}. \ISO~LWS data between 43--194\,\micron\ are
729: presented by \citet{cox1996}. Other data are taken from
730: \citet{sandell1994}, \citet{jenness2002}, and \citet{omont1995}.
731: \label{sed:crl2688}
732: }
733: \end{center}
734: \end{figure}
735:
736:
737: \subsection{CRL~2688} \label{crl}
738: CRL~2688 is a post-asymptotic giant-branch star with a proto-planetary
739: nebula \citep[``The Egg''][]{price1983}, and it is considered an
740: archetypal object with which to study the evolution of AGB red giants
741: into bipolar planetary nebulae. The distance to CRL~2688 is not
742: well-known, but is estimated to be approximately $1.25\,$kpc
743: \citep{crampton1975,cohen1977}. Despite residing in the Galactic
744: plane, CRL~2688 exhibits compact submillimeter emission and sufficient
745: contrast relative to its surroundings to provide a commonly used
746: secondary calibrator \citep[e.g.][]{sandell1994,jenness2002}. The
747: submillimeter emission arises from a dust shell with an extent of
748: approximately 5\arcsec\ \citep{jenness2002}, making it point-like for
749: BLAST. CRL~2688 was visited regularly throughout the flight and thus
750: the observations were useful for tracking the PSF shape (Fig.~2) and
751: relative antenna gain variations, making it our primary flat-fielding
752: calibrator.
753:
754: Fainter surface-brightness shells around CRL~2688 have also been
755: observed at 120--180\,\micron with \ISO\ at a radial distance
756: of $\sim$\,150\arcsec\ and 300\arcsec\ \citep{speck2000},
757: although this has recently been disputed with \Spitzer\ MIPS
758: observations \citep{do2005}. In any case, these shells would be large
759: compared to the BLAST beam, and of uniform surface-brightness, such
760: that any faint emission would be removed from the measured fluxes
761: through our baseline subtraction. No long-term submillimeter
762: variability has been detected in submillimeter observations with SCUBA
763: \citep{jenness2002} and hence archival data can be combined to perform
764: our calibrations.
765:
766: The mid-IR to FIR SED has been accurately measured by \ISO~LWS
767: between 43 and 194\,\micron\ \citep{cox1996}. At submillimeter
768: wavelengths, \citet{jenness2002} report the integrated fluxes
769: derived from SCUBA maps at 450 and 850$\mu$m, over a 40$''$ aperture
770: which exceed the earlier single-beam peak fluxes of Sandell
771: (1994). Combining these data with the new BLAST measurements
772: (Figure~\ref{sed:crl2688}) we find an acceptable fit to a
773: single-temperature modified black-body model, deriving a dust-temperature of
774: 210\,K and emissivity-index of $0.4 \pm 0.2$. The SED of CRL~2688 in the
775: FIR--submillimeter regime traced by BLAST is not as accurately
776: constrained as that of Arp~220 as there are no FIR measurments with either
777: IRAS or \Spitzer\ MIPS due to saturation. We derive uncertainties in the BLAST fluxes of
778: CRL~2688 that range from $\sim$\,10\% at 500\,\micron\ to $\sim$\,20\%
779: at 250\,\micron.
780:
781: \begin{figure}[t]
782: \begin{center}
783: \includegraphics[width=3in]{f4.eps}
784: \caption{Pallas spectral energy distribution. Symbols and lines are as
785: those described in Fig.~\ref{sed:arp220}.}
786: \label{sed:pallas}
787: \end{center}
788: \end{figure}
789:
790: \subsection{2~Pallas}
791: 2~Pallas is the third largest asteroid, with a diameter of 530\,km
792: \citep{mitchell1996}. It has a well-determined, eccentric orbit with a
793: semi-major axis between 2.1 and 3.4\,AU. At the time of the BLAST
794: observations, 2~Pallas was at a distance of 2.2\,AU\@. With an
795: angular size ${<}\,1$\arcsec, Pallas is point-like to BLAST\@. Errors
796: in the derived flux-densities of Pallas at BLAST wavelengths are
797: 5--10\%, and are due mostly to uncertainties in the size, shape, and
798: albedo of the asteroid. The current best estimates for the physical
799: properties of Pallas come from other submillimeter and IR flux density
800: measurements integrated into the ``thermo-physical'' model (TPM) of
801: \citet{muller2002}. The TPM has an uncertainty of around 5--10\% for
802: integrated fluxes in the {\it ISO\/} LWS band. The absolute
803: calibration of the model has a supposed accuracy of ${\sim}\,$10\% in the
804: longer-wavelength BLAST bands, comparing predictions with measurements
805: from the JCMT (850\,\micron) and CSO (350\,\micron) (T.~M{\"u}ller,
806: private communication). Although this uncertainty is similar to that for the
807: SED of Arp~220, Arp~220 was adopted as the absolute flux
808: calibrator over Pallas since it is brighter, and its SED does not vary
809: over the duration of the flight. Changes in apparent brightness
810: due to orbital effects are negligible on the timescale of the
811: individual BLAST observations \citep{muller2005}, and the variations
812: due to the 7.8~hour rotational period of 2~Pallas are less than
813: 7\% during the entire BLAST flight (T.~M{\"u}ller, private
814: communication). Although the TPM contains a full shape description,
815: and spin model (enabling SED predictions at any epoch), the
816: uncertainty in the absolute calibration of the model flux is
817: comparable to the predicted maximal brightness variations during the
818: BLAST flight. For this reason, in Table~2 we have simply averaged
819: together our 4 sets of Pallas observations, and we plot only the BLAST
820: photometry points in Figure~\ref{sed:pallas}, with no attempt to
821: relate these to other data at significantly different epochs.
822: %
823: The TPM predicts Pallas fluxes
824: of 13.2, 6.8, and 3.4 Jy which are consistant with the BLAST05
825: measurments, differing by 14, 8.6, and 7.3\% at 250, 350, and
826: 500\,\micron\ respectively. The submillimeter spectral-index
827: $\beta$ in the SED of
828: Pallas, as measured by BLAST, is $0.2 \pm 0.3$, which is the
829: shallowest of all the sources described in this paper.
830:
831:
832: \begin{figure}[t]
833: \begin{center}
834: \includegraphics[width=3in]{f5.eps}
835: \caption{L1014-IRS spectral energy distribution. Symbols and lines
836: are as those described in Fig.~\ref{sed:arp220}. The fit is poor,
837: partly because a single temperature modified blackbody may not be the best
838: model, but also because of resolution issues with some of the
839: photometry. We have not included the lower 450 and 850\,\micron\ data
840: points, which clearly disagree with the remaining photometry, probably
841: because of resolution effects. Non-BLAST data are from
842: \citet{visser2002} and \citet{young2004}.
843: \label{sed:l1014}
844: }
845: \end{center}
846: \end{figure}
847:
848:
849: \subsection{L1014-IRS}
850: LDN~1014 \citep{lynds1962}, which was classified as a ``starless
851: core'' due to the lack of a FIR \IRAS\ detection and no signature of a
852: continuum outflow \citep{young2004}, is a nearby dark cloud that
853: contains one of the lowest luminosity proto-stellar systems, L1014-IRS.
854: More recently, a bipolar molecular outflow has been discovered
855: \citep{bourke2005}. The distance to LDN~1014 is not firmly determined,
856: with estimates of $\sim$200\,pc \citep{huard2006}, 400--900\,pc
857: \citep{morita2006} and $< 500$\,pc \citep{shirley2007}.
858: The FIR--millimeter SED of L1014-IRS, derived from a physical
859: model of a circumstellar disk heated by the central protostellar
860: object, and constrained by recent \Spitzer\ observations from
861: 3--70\,\micron\ predict a strong spectral-peak at
862: $\sim$350\,\micron\ \citep{young2004}.
863: Since L1014--IRS was unresolved by BLAST, we were
864: unable to observe the predicted density-profile. The BLAST fluxes at
865: 350 and 500\,\micron, however, are in good agreement with the model, but
866: suggest that the SED peaks at shorter wavelengths, $< 250$\,\micron.
867:
868:
869: \begin{figure*}[t]
870: \begin{center}
871: \includegraphics[width=2.0in]{f6a.eps}
872: \includegraphics[width=2.2in]{f6b.eps}
873: \includegraphics[width=2.2in]{f6c.eps}
874: \caption{Left: 250\,\micron\ image of the K3-50 region, with the
875: brighter source K3-50A to the south-west and K3-50C to the north-east.
876: Spectral energy distributions for K3-50A (middle panel) and K3-50C
877: (right panel). Symbols and lines are as those described in
878: Fig.~\ref{sed:arp220}. Other data are taken from \citet{clegg1976}
879: and \citet{wynn-williams1977}. The resolution of \IRAS\ makes it
880: impractical to extract short wavelength flux densities for K3-50C.
881: \label{sed:k3-50}
882: }
883: \end{center}
884: \end{figure*}
885:
886:
887:
888: \begin{figure*}
889: \begin{center}
890: \includegraphics[width=2.8in]{f7a.eps}
891: \includegraphics[width=3in]{f7b.eps}
892: \includegraphics[width=3in]{f7c.eps}
893: \includegraphics[width=3in]{f7d.eps}
894: \caption{Top-left: 250\,\micron\ BLAST image of the W~75N region with
895: green contours showing the distribution of IR extinction \citep{kumar2007}.
896: We divide
897: the BLAST emission into 3 sources identified with: W~75N to the north;
898: DR21(OH) the northern part of the merged pair; and DR21, the
899: southernmost source. The 1\arcmin\ pixel-size in this map hides much
900: of the structure in the BLAST PSF.
901: SEDs show data and best-fitting models for W~75N (top-right),
902: DR21~(OH) (lower-left), and DR21 (lower-right). Symbols and lines are
903: as those described in Fig.~\ref{sed:arp220}. The data from
904: \citet{clegg1976}, \citet{davis2007}, \citet{helou1988}, and
905: \citet{harvey1977} are not adopted in the fits because of the smaller
906: aperture used to determine those flux density measurements. The data
907: from \citet{harvey1977} suggest a temperature of 65\,K for W~75N and
908: 40\,K for DR21~(OH), and these are used as constraints in our SED
909: fits. Other data are from \citet{clegg1976} and \citet{colome1995}.
910: \label{sed:w75n}
911: }
912: \end{center}
913: \end{figure*}
914:
915: \subsection{K3-50}
916: K3-50 is a group of compact \ion{H}{2} regions within the star-forming
917: complex W~58, at a distance of about 8.5\,kpc \citep{peeters2002}.
918: Despite the non-optimal PSF, BLAST clearly resolves at least 2 sources
919: separated by ${\simeq}$\,2\arcmin, which can be identified as K3-50A
920: and K3-50C \citep{howard1996}. Figure~\ref{sed:k3-50} shows the
921: 250\,\micron\ image. Fluxes for K3-50A and K3-50C are obtained by
922: simultaneously fitting to the BLAST data a model of two 2-dimensional
923: Gausssian sources convolved with the BLAST PSFs\@, whose sizes,
924: positions, and amplitudes are parameters in the fit. An alternative
925: model of two point-sources, fixed at the positions of K3-50A and
926: K3-50C \citep[from][]{howard1996}, produces an indistinguishable
927: solution for the inferred fluxes.
928:
929:
930: \citet{thompson2006} report an integrated $450\,\mu$m SCUBA flux of
931: $256\pm79\,$Jy for K3-50A, which is consistent with our $500\,\mu$m
932: photometry, despite issues with chopping and the restricted map-size
933: for SCUBA making the comparison potentially complicated. Although there
934: is some suggestion in the literature that these sources may be
935: variable\footnote{\url{http://www.jach.hawaii.edu/JCMT/continuum/calibration/sens/potentialcalibrators.html}},
936: we fit their individual FIR--millimeter SEDs (Figure~\ref{sed:k3-50}) with
937: single-temperature modified black-body emission. K3-50C appears to have a dust temperature
938: of only 9\,K, substantially colder than that in K3-50A which has
939: dust radiating at temperatures of $\sim$40\,K
940: (Table~\ref{fits}), in good agreement with previous estimates
941: based on FIR KAO data \citep{thronson1979}.
942: This differece in dust-temperature may be due to the fact that K3-50C
943: is embedded more deeply in the molecular cloud than K3-50A, or it may
944: reflect the different evolutionary stages of the objects, with K3-50A
945: generally agreed to be younger \citep{howard1996}.
946:
947:
948: \subsection{W~75N}
949: The W~75N \citep{westerhout1958} and DR21 field contains a set of
950: young protostar/compact \ion{H}{2} regions located in the Cygnus~X
951: molecular cloud complex. These sources are at a distance of about
952: 3\,kpc \citep{campbell1982,pipenbrink1988}, although other estimates
953: suggest a closer distance of around $1.7\,$kpc \citep{jakob2007}. The
954: region has been extensively studied over a wide wavelength range, with
955: many detailed spectroscopic (including OH and other maser lines) as
956: well as
957: continuum observations
958: \citep[see e.g.][]{davis2007}. Despite its extended emission,
959: W~75N has been suggested as a potential calibrator for submillimeter
960: observations \citep{sandell2003}. The W~75N region contains at
961: least 3 point-like objects which are resolved by BLAST: W~75N; DR21
962: \citep{downes1966}, and DR21(OH)\@. Flux densities for W~75N are
963: obtained by the matched filter technique, as described in
964: \S~\ref{sec:red}. Since DR21(OH) and DR21 overlap, their flux
965: densities are derived from a model of point sources convolved with the
966: BLAST PSF (as described in \S5.1), where the positions and fluxes are
967: the free parameters. These BLAST data suggest that DR21 has a
968: higher value of $\beta$ and a lower dust-temperature than DR21~(OH),
969: in agreement with what has been found in other studies
970: \citep[e.g.][]{jakob2007}.
971:
972: In addition to the detection of the protostellar cores within the \ion{H}{2}
973: regions, and as can be seen in Figure~\ref{sed:w75n}, the BLAST data
974: also show significant extended surface-brightness structure
975: ($\sim 1000$\,MJy\,sr$^{-1}$ at 250\,\micron). This is highly-correlated with the
976: near-IR-extinction maps towards the same region \citep{kumar2007}, indicating that it is real,
977: rather than being map-making artefacts. .
978:
979: \begin{figure}[t]
980: \begin{center}
981: \includegraphics[width=3in]{f8.eps}
982: \caption{ Mrk~231 spectral energy distribution showing data and
983: best-fitting model. Symbols and lines are as those described in
984: Fig.~\ref{sed:arp220}. Mrk~231 was not detected by BLAST at
985: 500\,\micron.
986: Other data are taken
987: from \citet{benford1999}, \citet{klaas2001}, \citet{lisenfeld2000}, \citet{stickel2004},
988: and \citet{helou1988}.
989: \label{sed:mrk231}
990: }
991: \end{center}
992: \end{figure}
993:
994: \subsection{Mrk~231}
995: Mrk~231 \citep{markarian1969}, also known as IRAS~12540+5708, is the
996: most luminous infrared galaxy within the local volume out to a
997: distance of 175\,Mpc. Mrk~231 hosts a Seyfert~1 nucleus, although the
998: FIR emission is known to be dominated by star-formation within a
999: region which is of order 1\arcmin\ \citep[see][and references
1000: therein]{lipari2005}. It is therefore point-like for BLAST, given
1001: the measured PSF (Fig.~1). Mrk~231 has a well measured SED and has
1002: been used as a template for photometric redshift estimates of extragalactic
1003: sources \citep[e.g.][]{klaas2001, aretxaga2005}. There are useful
1004: far-IR data from \IRAS\ and \ISO, although \citet{klaas2001} describe
1005: the existing {\it ISO\/} photometry as ``partly distorted and
1006: uncertain''. \citet{benford1999} reports a measurement at
1007: 350\,\micron\ of $1.05\pm0.25\,$Jy, which is consistent with the BLAST
1008: 350\,\micron\ photometry. There is also 450\,\micron\ photometry
1009: published, although using a much smaller aperture
1010: \citep{rigopoulou1996}. BLAST flux densities thus help constrain the
1011: SED in the poorly-sampled submillimeter regime (see
1012: Figure~\ref{sed:mrk231}). However, given the relative faintness of
1013: this source, the improvement is modest compared with other sources
1014: reported here. We find slightly lower values for temperature and
1015: $\beta$ compared with \citet{klaas2001}, but they also claim that the data
1016: are better fit with multiple temperature components.
1017:
1018: \begin{figure*}
1019: \includegraphics[width=2in]{f9a.eps}
1020: \includegraphics[width=2in]{f9b.eps}
1021: \includegraphics[width=2in]{f9c.eps}
1022: \includegraphics[width=2in]{f9d.eps}
1023: \includegraphics[width=2in]{f9e.eps}
1024: \caption{Spectral energy distributions of compact protostellar
1025: sources, including the BLAST fluxes (see Table \ref{calib}). The data symbols and
1026: fitted curves have the same meaning as those described in
1027: Fig. \ref{sed:l1014} -- (a) IRAS~20126+4104 --- Additional data are
1028: taken from \citet{helou1988}, \citet{hunter2000}, \citet{beuther2002},
1029: and \citet{williams2004}. The
1030: 350\,\micron\ data point from \citet{hunter2000} is not used in the
1031: fit, due to the smaller aperture. -- (b) IRAS~21078+5211 --- Additional
1032: data are taken from \citet{helou1988} and \citet{jenness1995}. Three
1033: data points are not used in the fit due to
1034: the small effective aperture sizes in those measurements. -- (c)
1035: IRAS~21307+5049 --- Additional data are taken from \citet{molinari2000}
1036: and \citet{helou1988}. The 850\,\micron\ data point from
1037: \citet{molinari2000} is not used in the fit due to its smaller
1038: aperture compared to the BLAST PSF. -- (d) IRAS~22134+5834 ---
1039: Additional data are taken from \citet{helou1988},
1040: \citet{chini2001}, \citet{beuther2002}, and \citet{williams2004}.
1041: -- (e) IRAS~23011+6126 --- Additional data are taken
1042: from \citet{helou1988} and \citet{chini2001}. The 850\,\micron\ data
1043: point from \citet{chini2001} is not used in the fit due to the smaller
1044: aperture. }
1045: \label{sed:protostellar}
1046: \end{figure*}
1047:
1048:
1049: \subsection{Compact Protostellar Sources} \label{sec:sources}
1050:
1051: The remaining Galactic targets, described below, are all compact
1052: protostellar sources, identified via colour-criteria in the FIR, that
1053: satisfy the BLAST selection requirements.
1054: We present here only their fluxes,
1055: dust masses and luminosities (Tables 2 \& 3) derived from the
1056: best-fit models describing their SEDs (Fig~\ref{sed:protostellar}).
1057:
1058: \subsubsection{IRAS~20126+4104}
1059: IRAS~20126+4104 is a high-mass compact protostar, located towards the
1060: Cygnus~X molecular cloud complex, although its exact distance is
1061: uncertain. We adopt a distance of $1.7\,$kpc \citep{shepherd2000}. It
1062: is probably the best studied example of a massive proto-stellar system
1063: associated with a disk and outflow \citep{edris2005}.
1064: BLAST
1065: provides better constraints on the SED \citep{wilking1989,hunter2000},
1066: particularly the submillimeter slope, as shown in
1067: Figure~\ref{sed:protostellar}.
1068:
1069: \subsubsection{IRAS~21078+5211}
1070: IRAS~21078+5211 is a compact \ion{H}{2} region with a water maser,
1071: located at a distance of approximately 1.65\,kpc
1072: \citep{molinari1996,wouterloot1993}.
1073: The best-fit
1074: single-temperature model with $T\,{\simeq}\,34\,$K and $\beta\,{\simeq}\,1.1$,
1075: is consistent with a previous study by ~\cite{mccutcheon1991}.
1076:
1077:
1078: \subsubsection{IRAS~21307+5049}
1079: IRAS~21307+5049 \citep[also Mol~136,][]{molinari1996} is a high mass
1080: protostellar candidate. At a distance of 3.6\,kpc, it has an
1081: angular size of about 4\arcsec\ in the millimeter \citep{molinari2002},
1082: and is therefore point-like
1083: to BLAST\@. \citet{fontani2004} present 850\micron\ SCUBA continuum
1084: observations, detecting an extended halo of 40\arcsec\ diameter and a
1085: compact unresolved core. The extended halo is not detected by BLAST\@.
1086: We only fit a
1087: single-temperature modified blackbody curve to the cold core
1088: component. We find a temperature of $32\,$K and luminosity of $L_{\rm
1089: FIR}\,{\simeq}\,4{,}000\,{\rm L}_\odot$, in very good agreement with
1090: \citet{molinari2000}.
1091:
1092: \subsubsection{IRAS~22134+5834}
1093: IRAS~22134+5834 is a medium-to-high mass protostar located within the
1094: extensive \ion{H}{2} region S134, in the constellation
1095: Cepheus. It is at a distance of approximately 900\,pc \citep{dobashi1994}, and
1096: has a total luminosity of ${\sim}\,1600\,{\rm L}_{\odot}$\@.
1097:
1098: \subsubsection{IRAS~23011+6126}
1099: IRAS~23011+6126 is a young protostar located at a distance of
1100: $\sim 730$\,pc \citep{crawford1970}. IRAS~23011+6126 is point-like to
1101: BLAST\@. \citet{chini2001} discusses the current best SED (similar to
1102: IRAS~22134+5834 above). The peak of the SED is at about 150\,\micron,
1103: as shown in Figure~\ref{sed:protostellar}. It appears to have
1104: $T\,{\simeq}\,28\,$K, with a shallower slope ($\beta = 0.6 \pm 0.2$)
1105: than most of the other protostellar sources that BLAST observed.
1106:
1107:
1108: \section{Conclusions}
1109: For the BLAST 2005 flight, the out-of-focus PSFs significantly reduced the
1110: point-source sensitivity, and yet because this is such a difficult waveband to
1111: observe from the ground, the BLAST05 data still produced useful results. The
1112: submillimeter flux densities reported here agree well with other measurements,
1113: at least when the aperture sizes are consistent. However, BLAST05 uncertainties are
1114: typically much smaller than those of other measurements in the
1115: 250--500\,\micron\ regime.
1116:
1117: A particular advantage of the BLAST data is the ability to constrain the
1118: emissivity index $\beta$, due primarily to the 3 separate submillimeter
1119: channels, observed simultaneously with the same telescope, and having a
1120: consistent calibration procedure performed across all 3 bands.
1121:
1122: Arp~220 was adopted as the primary calibrator for the BLAST05 data. A
1123: useful cross-check is to note that the derived flux densities for
1124: Pallas were all within $1\sigma$ of the values derived from the TPM
1125: \citep{muller2002}.
1126:
1127: By applying this careful calibration procedure, BLAST has been able to
1128: improve estimates of dust temperatures and masses for several
1129: relatively bright sources, and also to provide a database of
1130: submillimeter measurements which may be useful for calibrating future
1131: instruments.
1132:
1133: \acknowledgments
1134: The BLAST collaboration acknowledges the support of
1135: NASA through grant numbers NAG5-12785, NAG5-13301, and NNGO-6GI11G, the
1136: Canadian Space Agency (CSA), Canada's Natural Sciences and Engineering
1137: Research Council (NSERC), and the UK Particle Physics \& Astronomy
1138: Research Council (PPARC).
1139: We would also like to thank the Columbia Scientific Balloon Facility (CSBF)
1140: staff for their outstanding work, as well as
1141: T.G.~M{\"u}ller for valuable discussions on Pallas.
1142: LO acknowledges partial support by the Puerto Rico
1143: Space Grant Consortium and by the Fondo Istitucional
1144: para la Investigacion of the University of Puerto Rico.
1145: CBN acknowledges support from the Canadian Institute for Advanced Research.
1146: This research has been enabled by the use of WestGrid computing resources.
1147: This research also made use of the SIMBAD database, operated at the
1148: Centre de Don{\'e}es astronomiques de Strasbourg (CDS), Strasbourg, France,
1149: and the NASA/IPAC Extragalactic Database (NED), which is operated by the
1150: Jet Propulsion Laboratory, California Institute of
1151: Technology, under contract with the National
1152: Aeronautics and Space Administration.
1153:
1154: \bibliographystyle{apj}
1155: \bibliography{ms}
1156:
1157: \clearpage
1158:
1159: \begin{deluxetable}{cccccc}
1160: \tablecaption{Calibration Coefficients and Uncertainties for BLAST05\label{calib}}
1161: \tablewidth{0pt}
1162: \tablehead{
1163: \colhead{Band} & \colhead{calib.~coeff.} & \colhead{uncertainty} &
1164: \multicolumn{3}{c}{Pearson correlation matrix} \\
1165: \colhead{[\micron]} & \colhead{[$\times 10^{12}$ Jy V$^{-1}$]} & \colhead{[\%]} &
1166: \colhead{250\,\micron} & \colhead{350\,\micron} & \colhead{500\,\micron}
1167: }
1168: \startdata
1169: 250 & 7.61 & 12 & 1 & 0.97 & 0.87 \\
1170: 350 & 3.16 & 10 & & 1 & 0.96 \\
1171: 500 & 1.56 & 8 & & & 1 \\
1172: \enddata
1173: \tablecomments{Calibration coefficients, calibration uncertainties, and Pearson
1174: correlation matrix, showing the relationship between errors in different bands
1175: for BLAST05.}
1176: \end{deluxetable}
1177:
1178:
1179: \begin{deluxetable}{lccrrr}
1180: \tabletypesize{\scriptsize}
1181: %\rotate
1182: \tablecaption{Flux Densities of BLAST05 Targeted Sources\label{fluxen}}
1183: \tablewidth{0pt}
1184: \tablehead{
1185: & \colhead{RA\tablenotemark{a}} & \colhead{DEC\tablenotemark{a}} &
1186: \multicolumn{3}{c}{Flux Density [Jy]} \\
1187: \colhead{Name} & \colhead{[J2000]} & \colhead{[J2000]} &
1188: \colhead{250\,\micron} & \colhead{350\,\micron} & \colhead{500\,\micron}
1189: }
1190: \startdata
1191: Pallas\tablenotemark{b} & \nodata & \nodata &$11.6\pm1.5$&$ 6.3\pm0.7$&$ 3.7\pm0.3 $\\
1192: CRL~2688 & $21\h 02\m 18\fs75$ & $+36\degr 41\arcmin 37\farcs8$ &$ 113\pm14 $&$ 49\pm5 $&$21.6\pm1.7 $\\
1193: LDN~1014 & $21\h 24\m 06\s$ & $+49\degr 59\farcm1 $ &$22.4\pm2.7$&$16.9\pm1.7$&$ 8.9\pm0.7 $\\
1194: IRAS~20126+4104 & $20\h 14\m 25\fs1$ & $+41\degr 13\arcmin 32\arcsec$ &$ 590\pm71 $&$ 256\pm26 $&$ 97\pm8 $\\
1195: IRAS~21078+5211 & $21\h 09\m 25\fs2$ & $+52\degr 23\arcmin 44\arcsec$ &$1300\pm160$&$ 590\pm60 $&$ 260\pm21 $\\
1196: IRAS~21307+5049 & $21\h 32\m 31\fs5$ & $+51\degr 02\arcmin 22\arcsec$ &$ 102\pm12 $&$ 46\pm5 $&$ 18\pm1.5 $\\
1197: IRAS~22134+5834 & $22\h 15\m 09\fs1$ & $+58\degr 49\arcmin 09\arcsec$ &$ 418\pm50 $&$ 208\pm21 $&$ 82\pm7 $\\
1198: IRAS~23011+6126 & $23\h 03\m 13\fs9$ & $+61\degr 42\arcmin 21\arcsec$ &$ 95\pm12 $&$ 59\pm6 $&$ 31\pm2.5 $\\
1199: K3-50A\tablenotemark{c} & $20\h 01\m 45\fs6$ & $+33\degr 32\arcmin 42\arcsec$ &$2100\pm270$&$ 590\pm67 $&$ 270\pm26 $\\
1200: K3-50C\tablenotemark{c} & $20\h 01\m 54\fs2$ & $+33\degr 34\arcmin 15\arcsec$ &$1870\pm240$&$1000\pm110$&$ 370\pm35 $\\
1201: W~75N\tablenotemark{d} & $20\h 38\m 36\fs5$ & $+42\degr 37\arcmin 35\arcsec$ &$4500\pm540$&$2000\pm200$&$ 730\pm60 $\\
1202: DR21~(OH)\tablenotemark{d} & $20\h 39\m 00\fs9$ & $+42\degr 22\arcmin 38\arcsec$&$9100\pm1400$&$4600\pm660$&$2540\pm330 $\\
1203: DR21\tablenotemark{d} & $20\h 39\m 01\fs1$ & $+42\degr 19\arcmin 43\arcsec$&$8800\pm1400$&$3900\pm550$&$ 920\pm120 $\\
1204: Mrk~231 & $12\h 56\m 14\fs23$ & $+56\degr 52\arcmin 25\farcs2$ &$ 6.0\pm0.8$&$ 1.5\pm0.3$& \nodata\tablenotemark{e} \\
1205: Arp~220\tablenotemark{f} & $15\h 34\m 57\fs21$ & $+23\degr 30\arcmin 09\farcs5$ &$24.2 $&$ 9.8 $&$ 3.9 $
1206: \enddata
1207: \tablecomments{Flux densities and associated uncertainties for BLAST05 targeted
1208: sources in the 250, 350, and 500\,\micron\ bands. Quoted flux densities
1209: have been color-corrected. The uncertainties include the estimated contributions
1210: from calibration uncertainty as well as instrumental noise.}
1211: \tablenotetext{a}{Positions are nominal as given by SIMBAD.}
1212: \tablenotetext{b}{Pallas flux densities are average values from 4 observations
1213: taken at the following Julian Dates: 2453534.26; 2453534.29; 2453535.22; and
1214: 2453535.40.}
1215: \tablenotetext{c}{The K3-50 region is resolved into two sources by BLAST05.}
1216: \tablenotetext{d}{The W~75N region is resolved into three sources by BLAST05.}
1217: \tablenotetext{e}{Mrk~231 was not detected at 500\,\micron\ by BLAST05.}
1218: \tablenotetext{f}{Arp~220 is the absolute flux calibrator for BLAST05.
1219: Flux densities presented here should thus be considered as
1220: predictions based on a model fit to data from other instruments.
1221: Uncertainties in this model are given in \S~\ref{sec:arp220}}.
1222: \end{deluxetable}
1223:
1224: \begin{deluxetable}{lrrrrrrr}
1225: \tabletypesize{\scriptsize}
1226: \tablecaption{Single Temperature SED Best Fit Parameters of BLAST05 Targeted Sources\label{fits}}
1227: \tablewidth{0pt}
1228: \tablehead{
1229: \colhead{Name} & \colhead{$T$} & \colhead{$\beta$} & \colhead{$S_{\rm FIR}$} &
1230: \colhead{Distance} & \colhead{$L_{\rm FIR}$} & \colhead{$M_{\rm dust}$} \\
1231: & \colhead{[K]} & & \colhead{[W m$^{-2}$]} & \colhead{[kpc]} & \colhead{[L$_\odot$]}
1232: & \colhead{[M$_\odot$]}
1233: }
1234: \startdata
1235: Pallas\tablenotemark{a} & $50.2$ & $0.2 \pm 0.2$ & $1.2\E{-12}$ &
1236: \nodata & \nodata & \nodata \\
1237: CRL~2688\tablenotemark{a} & $210$ & $0.4 \pm 0.2$ & $1.2\E{-9}$ &
1238: 1.25 & $1.7\E{5}$ & $3.2\E{-3}$ \\
1239: LDN~1014 & $12 \pm 3$ & $1.8 \pm 0.5$ & $(3.3 \pm 1.2)\E{-13}$ &
1240: 0.2 & $0.38 \pm 0.14$ & $(3.7 \pm 3.5)\E{-3}$ \\
1241: IRAS~20126+4104\tablenotemark{a} & $41$ & $1.1 \pm 0.2$ & $9.3\E{-11}$ &
1242: 1.7 & $4.0\E4$ & $0.67$ \\
1243: IRAS~21078+5211 & $33.7 \pm 2.0$ & $1.0 \pm 0.2$ & $(1.2 \pm 0.1)\E{-10}$ &
1244: 1.65 & $(1.0 \pm 0.1)\E4$ & $0.29 \pm 0.05$ \\
1245: IRAS~21307+5049 & $32.9 \pm 2.5 $ & $1.2 \pm 0.2$ & $(1.0 \pm 0.1)\E{-11}$ &
1246: 3.6 & $(4.3 \pm 0.5)\E3$ & $0.11 \pm 0.03$ \\
1247: IRAS~22134+5834 & $23.5 \pm 1.0 $ & $2.1 \pm 0.4$ & $(2.0 \pm 0.8)\E{-11}$ &
1248: 0.90 & $520 \pm200$ & $0.11 \pm 0.07$ \\
1249: IRAS~23011+6126 & $27.8 \pm 2.8 $ & $0.6 \pm 0.2$ & $(4.3 \pm 0.5)\E{-12}$ &
1250: 0.73 & $71 \pm 9$ & $0.01 \pm 0.007$\\
1251: K3-50A & $39.7 \pm 2.8$ & $1.7 \pm 0.2$ & $(6.2 \pm 0.5)\E{-10}$ &
1252: 8.5 & $(1.4 \pm 0.1)\E6$ & $8.1 \pm 1.4$ \\
1253: K3-50C & $9.3 \pm 1.6$ & $3.9 \pm 0.5$ & $(2.3 \pm 0.8)\E{-11}$ &
1254: 8.5 & $(5.2 \pm 2.0)\E4$ & $(1.2 \pm 1.1)\E3$ \\
1255: W~75N & $65$\tablenotemark{b} & $0.8 \pm 0.2$ & $(2.5 \pm 1.3)\E{-9}$ &
1256: 3 & $(7.1 \pm 3.6)\E{5}$ & $1.0 \pm 0.1$ \\
1257: DR21~(OH) & $40$\tablenotemark{c} & $0.5 \pm 0.3$ & $(7.3 \pm 3.3)\E{-10}$ &
1258: 3 & $(2.1 \pm 0.9)\E5$ & $4.9 \pm 0.7$ \\
1259: DR21 & $29.2\pm3.0$ & $2.3 \pm 0.2$ & $(1.7 \pm 0.3)\E{-9}$ &
1260: 3 & $(4.9 \pm 0.8)\E5$ & $8.7 \pm 3.0$ \\
1261: Mrk~231 & $43.9 \pm 2.5$ & $1.5 \pm 0.1$ & $(1.8 \pm 0.2)\E{-12}$ &
1262: 175\,000 & $(1.6 \pm 0.1)\E{12}$ & $(6.4 \pm 0.8)\E{6}$\\
1263: Arp~220\tablenotemark{d}
1264: & $41.7 \pm 3.5$ & $1.3 \pm 0.1$ & $(5.9 \pm 0.8)\E{-12}$ &
1265: 75\,000 & $(1.04 \pm 0.14)\E{12}$ & $(7.8 \pm 1.9)\E6$
1266: \enddata
1267: \tablecomments{Parameters for a single temperature modified blackbody fit to
1268: the SED.
1269: $T$ and $\beta$ are the best fit to the BLAST05 and other data as indicated
1270: in the text. $S_{\rm FIR}$ is the total flux from the modified blackbody fit.
1271: Errors are from 100 Monte Carlo simulations of the fit. Distances given
1272: are from references indicated in \S~\ref{sec:obs}. FIR luminosity is
1273: based on these adopted distances. $M_{\rm dust}$ is based on the formula
1274: $M_{\rm dust} = (S_\nu D^2)/(\kappa B_\nu(T))$,
1275: with an assumed value of 10 for $\kappa$ \citep[see][]{chapin2008}.}
1276: \tablenotetext{a}{The data included in the fit do not accurately constrain the
1277: temperature, so no error bars are given.}
1278: \tablenotetext{b}{Data from \citet{harvey1977} suggest a temperature of 65\,K,
1279: which is used as a constraint in the fit.}
1280: \tablenotetext{c}{Data from \citet{harvey1977} suggest a temperature of 40\,K,
1281: which is used as a constraint in the fit.}
1282: \tablenotetext{d}{Arp~200 is the primary calibrator for BLAST05. This is our
1283: best-fit to the other data, which we use to calibrate the BLAST photometry.}
1284: \end{deluxetable}
1285:
1286: \end{document}
1287: