The Astrophysical Journal Supplement Series, 154:259–265, 2004 September
# 2004. The American Astronomical Society. All rights reserved. Printed in U.S.A.
ENERGY SOURCES OF THE FAR-INFRARED EMISSION OF M33
J. L. Hinz, G. H. Rieke, K. D. Gordon, P. G. Pérez-González, C. W. Engelbracht, A. Alonso-Herrero,
J. E. Morrison, K. Misselt, and D. C. Hines1
Steward Observatory, University of Arizona, 933 North Cherry Avenue, Tucson, AZ 85721; jhinz@as.arizona.edu, grieke@as.arizona.edu,
kgordon@as.arizona.edu, pgperez@as.arizona.edu, cengelbracht@as.arizona.edu, aalonso@as.arizona.edu,
jmorrison@as.arizona.edu, kmisselt@as.arizona.edu, dhines@as.arizona.edu
R. D. Gehrz, E. Polomski, C. E. Woodward, and R. M. Humphreys
Department of Astronomy, University of Minnesota, Minneapolis, MN 55455; gehrz@astro.umn.edu, elwood@astro.umn.edu,
chelsea@astro.umn.edu, roberta@astro.umn.edu
M. W. Regan
Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218; mregan@stsci.edu
J. Rho
Spitzer Science Center, California Institute of Technology, MC 220-6, Pasadena, CA 91125; rho@ipac.caltech.edu
and
J. W. Beeman and E. E. Haller
Lawrence Berkeley National Laboratories, 1 Cyclotron Road, Berkeley, CA 94720
Receivv ed 2004 April 7; accepted 2004 May 12
ABSTRACT
We present observations of the spiral galaxy M33 with Spitzer at 24, 70, and 160 m. The excellent resolution
and mapping capabilities of Spitzer combined with the proximity of M33 result in observations that enable a
detailed study of the distribution of star formation (SF) and dust in the galaxy. We compare the morphology of
M33 at far-infrared wavelengths with other standard SF indicators such as H and radio continuum using a
Fourier filtering technique to separate the diffuse emission components from compact sources. We find that the
infrared emission at 24 and 70 m follows closely the structure of the ionized gas, indicating that it is heated
largely by hot, ionizing stars. At 160 m a diffuse cold dust component matches only approximately the structure
of the old red stellar population or the distribution of blue light. It is, however, very similar to the structure of the
diffuse nonthermal radio emission.
Subject heading
gs: galaxies: individual (M33) — galaxies: spiral — galaxies: structure — infrared: galaxies
1. INTRODUCTION
spiral (SA(s)cd) galaxy at a distance of 840 kpc (Freedman
et al. 1991). Its low metallicity (e.g., Cohen et al. 1984), bright
nucleus with a small bulge, and lack of a supermassive black
hole (Merritt et al. 2001; Gebhardt et al. 2001) set it apart from
our own and other nearby galaxies. For the current study, it
is sufficiently close for good physical resolution in the FIR
(100 4 pc) and nearly face-on (i ¼ 56 ; Regan & Vogel 1994),
making it virtually unique in the quality of comparisons
possible in the morphologies of different components of the
galaxy. As a result, M33 has been the subject of a number of
previous studies, but they have not achieved a consensus on
the physical processes causing the infrared emission. For example, Deul (1989) concludes from the morphology of the
Infrared Astronomical Satellite (IRAS ) images that the FIR
output is divided approximately equally between the H ii
regions and the diffuse interstellar radiation field. Walterbos &
Greenawalt (1996) model the IRAS radial distributions and
find that nearly all of the output can arise from the diffuse
interstellar field, but Devereux et al. (1997) find that most of
the energy is produced in the H ii regions.
The IRAS beam of 90 00 ; 4 0 (at 60 m) translates into about
400 ; 1000 pc at the distance of M33, large enough to make
the interpretation of the morphology ambiguous. Recently,
Hippelein et al. (2003) reported imaging with ISOPHOT using
a 6000 beam (translating to a diameter of about 250 pc), a
substantial improvement, i.e., a reduction in beam area of
Far-infrared ( FIR) fluxes from normal galaxies are commonly taken to indicate the rate of recent star formation.
However, placing this interpretation on firm footing has been
frustrated by uncertainties in the basic issue of what component of the galaxy provides the energy that is absorbed by the
dust and reradiated (Kennicutt 1998). For example, Devereux
and coworkers (Devereux et al. 1994, 1996, 1997; Devereux
& Scowen 1994) and Jones et al. (2002) argue from the close
correspondence between hydrogen recombination line emission and FIR morphologies that the FIR is powered predominantly by H ii regions. Deul (1989), Walterbos & Greenawalt
(1996), and Hirashita et al. (2003) argue that about half of the
FIR emission, perhaps more, is due to dust heated by a diffuse
‘‘interstellar radiation field,’’ defined as an average radiation
field that is not dominated by any particular star or star cluster
and that is of similar color to that of the solar neighborhood.
Sauvage & Thuan (1992) have suggested that the relative role
of young stars compared with the diffuse interstellar radiation
field increases with later galaxy type.
M33 (NGC 598), also known as the Triangulum galaxy, is
ideal for studying this question. It is a well-known late-type
1
Current address: Space Science Institute, 4750 Walnut Street, Suite 205,
Boulder, CO 80301.
259
260
HINZ ET AL.
Vol. 154
nearly an order of magnitude. They report two main dust
components: a warm (45 K) component that is present in the
spiral arms and star-forming regions, and a cold (16 K)
component that is distributed across the disk of the galaxy and
is presumably heated by diffuse interstellar radiation, as well
as localized cold dust associated with molecular clouds.
With the Multiband Imaging Photometer Spitzer for (MIPS),
the point-spread function is critically sampled, allowing another order-of-magnitude reduction in beam area. We report
new images of M33 at resolutions of 600 (25 pc) at 24 m,
1800 (75 pc) at 70 m, and 4000 (165 pc) at 160 m. These
images are compared with each other and with images of
potential energy sources for the FIR emission to identify how
the dust seen at these wavelengths is heated.
2. OBSERVATIONS AND DATA REDUCTION
Multiple overlapping medium-rate scans across the face of
the galaxy with MIPS produced maps at 24, 70, and 160 m
(see Engelbracht et al. 2004 in this volume for details on
MIPS large-galaxy observations). The final mosaics are shown
in Figure 1 ( Plate 1). Data reduction was performed using the
MIPS instrument team Data Analysis Tool (Gordon et al.
2004a). The 70 and 160 m images exhibit faint linear stripes
along the scan direction that are residual instrumental artifacts
due to the time-dependent responsivity of the Ge: Ga detectors.
The integrated flux in our 160 m image is 2054 411 Jy,
in good agreement with the value of 2200 Jy at 175 m
reported by Hippelein et al. (2003). Both measurements are
significantly brighter than the whole-galaxy measurement
from COBE reported by Odenwald et al. (1998). Hippelein
et al. (2003) suggest that the difference arises because of the
structure of the Galactic cirrus around the galaxy. We have
recomputed the DIRBE flux densities using an online photometry tool ,2 finding much better agreement with our data.
Figure 2 shows a spectral energy distribution (SED) plot for
M33 using data from MIPS, IRAS, and DIRBE. The total
emission in the MIPS images was measured using IRAF’s
polyphot task, while the IRAS measurements were taken from
Rice et al. (1988).
A variety of ancillary data have been collected to compare
the FIR structure of M33 with other images at wavelengths
indicative of the potential energy sources for heating the dust.
To locate the ionizing stars, we used H and 6 cm radio
continuum images from Devereux et al. (1997). The H
observations were obtained at the Case Western Burrell
Schmidt Telescope at Kitt Peak, providing a 1 field of view
with 200 pixels. The 6 cm observations were taken with the
Very Large Array (VLA) and the Westerbork Synthesis Radio
Telescope (WSRT). The typical sensitivity is 30 Jy. Their
clean radio mosaic was made by smoothing to a common
resolution of 1000, and all nonthermal sources were subtracted
as detailed in Devereux et al. (1997). The near-infrared
K-band image of M33 is taken from the Two Micron All Sky
Survey (2MASS) Large Galaxy Atlas (Jarrett et al. 2003). The
B-band images were taken at the Bok 2.3 m telescope on Kitt
Peak with the 90-Prime instrument (Williams et al. 2001, and
2004, in preparation) on 2003 December 29 during nonphotometric conditions. Six pointings were taken, with each
pointing consisting of at least four 1 minute exposures. The
mean sky level was subtracted from each of the six images
before combining into a single mosaic.
2
Available at http://lambda.gsfc.nasa.gov/product/cobe/ browser.cfm.
Fig. 2.—Infrared SED of M33.
For the purposes of comparison, the H and 24 m M33
images were convolved to the resolution of the MIPS 70 m
map, and all were cropped to a common field of view. The
convolution technique is described in detail by Engelbracht
et al. (2004). The convolved H and 24 m images, the Band K-band images, and the 70 and 160 m maps are shown in
Figure 3 ( Plate 2).
3. RESULTS
3.1. General Description of the Imag
ges
The near-IR image traces the smooth distribution of the
older stellar population in two broad and indistinct arms and
a small bulge, with diffuse emission across the disk of the
galaxy. The blue image is dominated by the spiral arms. The
H , 24, and 70 m and radio images also show distinct spiral
arms, studded with many individual H ii regions, and a near
absence of a bulge. These spiral arms appear more subdued
at 160 m and with smoother structure, indicating that the
individual H ii regions contribute less to the overall emission.
The H traces the individual sources well, dominated by
emission from H ii regions with a total luminosity of
7:06 1:40 ; 106 L (Devereux et al. 1997). As shown by
the smoother spiral arms in the 24 m image, the H emission
suffers from slight attenuation by dust in the star-forming
regions. For instance, in the northeastern spiral arm, just south
of the bright H ii region NGC 604, the 24 m image reveals a
stream of connected emission, while H shows only individual sources.
3.2. Comparison of Infrared and Compact Free-Free
Emission Reg
gions
Figure 4 shows the 6 cm radio map and a subtraction of the
6 cm map from the convolved 24 m image, performed using
an IRAF script originally designed to identify supernovae in
nearby galaxies (Van Dyk et al. 2000). Using input coordinates of identical objects in the two frames, the script fits
Gaussian profiles to those objects to get more accurate centers,
aligns the images, calculates the FWHM difference between
the two images (if any) and transforms one to another, calculates a scaling factor determined by the number of counts in
a user defined aperture around the chosen objects, and finally
subtracts the two images. Five bright H ii regions were used to
align the 24 m and 6 cm images. The resultant image reveals
almost no diffuse residual, except for false faint structure near
No. 1, 2004
ENERGY SOURCES OF FIR EMISSION OF M33
261
Fig. 4.—The 6 cm radio map from Devereux et al. (1997) is on the left. Subtraction of the 6 cm radio data from the 24 m map produces the image on the right.
North is up, and east is to the left. Both the 24 m and 6 cm data were convolved to the 70 m resolution of 1800 before subtraction.
the center of the galaxy caused by the dark noise patches in the
radio image. We conclude that the compact structures at 24 m
are heated almost entirely within the H ii regions. Because of the
close similarity in structure of the galaxy at 70 m, the compact
emission at the latter wavelength probably has the same origin.
At 160 m, the compact structure is subdued; some of the
emission must originate in the H ii regions, but it appears that
much of it may be more extended. We see no evidence for
sources that appear to be relatively brighter at 70 than at 24 m
as seen in other galaxies observed by Spitzer (e.g., M81;
Gordon et al. 2004b) or for increased extended emission at the
outskirts of the galaxy at the longer wavelengths.
3.3. Diffuse Emission
The images of M33 are dominated by the compact structures, which must be suppressed to study the diffuse emission.
Figure 5 shows images on which a Fourier transform has been
performed to filter out the high-frequency information. These
were created using the Perl Data Language fast Fourier
transform package where the boundary for the high-frequency
filter was chosen such that all point sources and structures
were suppressed, leaving only the diffuse emission. This corresponded to suppressing structures with frequencies smaller
than five 160 m pixels (300 pc).
In this comparison, we see that the H , 24, and 70 m
emission still look very similar, with the center of the galaxy
and NGC 604 being the prominent features, surrounded by
diffuse emission of virtually identical structure. Thus, even the
extended emission at these wavelengths appears to be powered largely by ionizing stars. The 160 m filtered map, on the
other hand, shows smooth contours of emission originating at
the center of M33 and a barlike structure elongated in the
northeast-southwest direction. This is roughly consistent with
the results of Hippelein et al. (2003), who observed diffuse
cold dust emission with a ‘‘disklike,’’ nonspiral structure in a
scaled difference map of 60 and 170 m images.
It has often been suggested that diffuse interstellar radiation
field heats the cold, diffuse dust. To test this hypothesis, we
have applied the same Fourier filtering technique to images at
K and B bands to test whether the cool stars, or hot but
nonionizing ones, have a similar distribution to the 160 m
diffuse image. As shown in Figure 5, the match to the B band
is poor—while there is some emission both in the B band and
at 160 m in the innermost regions of M33, the outer structures in the B band do not resemble the simple bar at 160 m.
The K-band image is largely centrally concentrated but does
seem to mirror the northeast-southwest elongation of the 160 m
emission. Thus, it appears that the cool and hot nonionizing
stars could contribute to the diffuse cold dust heating, particularly in the center of the galaxy. However, because the
K-band and 160 m smoothed images are not identical, it is
possible that some other component plays a role in heating
the dust.
Simple inspection of soft X-ray images of M33 (Haberl &
Pietsch 2001) shows that the diffuse emission is not matched
to the 160 m structure, with no elongation in the northeastsouthwest direction, implying that the hot thermal plasma is
not a source of cold dust heating. An elongated structure has
been seen in radio continuum measurements at 17.4 cm by
Buczilowski (1988). Figure 6 shows contour maps of the
distribution of the total intensity at 17.4 cm after subtraction of
24 unrelated sources (Buczilowski 1988). The thermal emission at 17.4 cm is estimated to be 15% (Buczilowski 1988)
of the total radio continuum emission. The compact structures
that dominate most radio maps stand out because the low
spatial frequencies are suppressed by interferometer-type
telescopes; for example, Viallefond et al. (1986) show that the
WSRT map at 1.4 GHz accounts for only 16% of the total
emission. Thus, the filled aperture image in Figure 6 represents almost entirely the diffuse nonthermal emission.
Figure 6 also shows the 160 m Fourier-filtered image
observed by MIPS. The image is very similar to the diffuse
nonthermal image. We conclude that the diffuse, cold dust is
heated not only by the old stellar population, but also by a
mechanism closely related to the nonthermal emission, either
directly by cosmic-ray heating of the interstellar grains or
Fig. 5.—Images shown in Fig. 3 have now been Fourier-filtered and are presented here at (a) H, (b) B band, (c) K band, (d) 24 m, (e) 70 m, and ( f ) 160 m.
North is up, and east is to the left. The field of view is approximately 38 0 ; 44 0.
Fig. 6.—Contour maps of the 17.4 cm data taken from (a) Fig. 1 of Buczilowski (1988) and the (b) 160 m Fourier-filtered image scaled to the same size.
ENERGY SOURCES OF FIR EMISSION OF M33
indirectly by some mechanism, for example, related to the
acceleration of the radio-emitting electrons that in parallel
heats the dust. A correlation between the nonthermal radio
emission and FIR emission has also been seen in M31
(Hoernes et al. 1998), where it is proposed that the correlation
arises due to a coupling of the magnetic field to the gas which
is mixed with the cool dust.
4. CONCLUSIONS
New MIPS images of M33 offer an unprecedented look at
the mechanisms for heating the dust that accounts for the FIR
emission of this well-studied late-type spiral. A morphological
comparison of the FIR data with ground-based H , near-IR,
and high-frequency radio continuum observations suggests
that the dust heating at 24 and 70 m is dominated by ionizing
stars. However, at 160 m, there is a diffuse, cold dust component with a dramatically different spatial distribution that
263
most closely matches the distribution of diffuse nonthermal
radio emission. Future work should address whether this cold
dust is heated directly by cosmic rays or whether it is heated in
parallel to cosmic-ray acceleration.
J. L. H. thanks Nick Devereux for kindly providing the
radio and H data sets. J. L. H. also thanks Grant Williams,
Ed Olszewski, and the 90-Prime Camera Team for allowing us
to use their B-band images of M33. This work is based on
observations made with the Spitzer Space Telescope, which is
operated by the Jet Propulsion Laboratory, California Institute
of Technology under NASA contract 1407. Support for this
work was provided by NASA through contract 960785, issued
by JPL/Caltech. R. D. G., E. P., C. E. W., and R. M. H. are
supported by NASA through the Spitzer GTO Program.
REFERENCES
Buczilowski, U. R. 1988, A&A, 205, 29
Jarrett, T. H., Chester, T., Cutri, R., Schneider, S. E., & Huchra, J. P. 2003, AJ,
Cohen, J. G., Persson, S. E., & Searle, L. 1984, ApJ, 281, 141
125, 525
Deul, E. R. 1989, A&A, 218, 78
Jones, L. V., Elston, R., & Hunter, D. 2002, AJ, 124, 2548
Devereux, N., Duric, N., & Scowen, P. A. 1997, AJ, 113, 236
Kennicutt, R. C. 1998, ARA&A, 36, 189
Devereux, N., Jacoby, G., & Ciadullo, R. 1996, AJ, 111, 2115
Odenwald, S., Newmark, J., & Smoot, G. 1998, ApJ, 500, 554
Devereux, N., Price, R., Wells, L. A., & Duric, N. 1994, AJ, 108, 1667
Merritt, D., Ferrarese, L., & Joseph, C. L. 2001, Science, 293, 1116
Devereux, N., & Scowen, P. A. 1994, AJ, 108, 1244
Regan, M. W., & Vogel, S. N. 1994, ApJ, 434, 536
Engelbracht, C., et al. 2004, ApJS, 154, 248
Rice, W., Lonsdale, C. J., Soifer, B. T., Neugebauer, G., Koplan, E. L., Lloyd,
Freedman, W. L., Wilson, C. D., & Madore, B. F. 1991, ApJ, 372, 455
L. A., de Jong, T., & Habing, H. J. 1988, ApJS, 68, 91
Gebhardt, K., et al. 2001, AJ, 122, 2469
Sauvage, M., & Thuan, T. X. 1992, ApJ, 396, L69
Gordon, K., et al. 2004a, PASP, submitted
Van Dyk, S. D., Peng, C. Y., King, J. Y., Filippenko, A. V., Treffers, R. R.,
———. 2004b, ApJS, 154, 215
Li, W., & Richmond, M. W. 2000, PASP, 112, 1532
Haberl, F., & Pietsch, W. 2001, A&A, 373, 438
Viallefond, F., Goss, W. M., van der Hulst, J. M., & Crane, P. C. 1986, A&AS,
Hippelein, H., Haas, M., Tuffs, R. J., Lemke, D., Stickel, M., Klaas, U., &
64, 237
Völk, H. J. 2003, A&A, 407, 137
Walterbos, R. A. M., & Greenawalt, B. 1996, ApJ, 460, 696
Hirashita, H., Buat, V., & Inoue, A. K. 2003, A&A, 410, 83
Williams, G. G., et al. 2001, BAAS, 33, 791
Hoernes, P., Berkhuijsen, E. M., & Xu, C. 1998, A&A, 334, 57
Fig. 1.—MIPS images of M33 at (a) 24, (b) 70 m, and (c) 160 m, for which the resolutions are 600 , 1800 , and 4000 , respectively. The field of view is
approximately 38 0 ; 40 0. North is up, and east is to the left. The coordinate system is J2000.0.
Plate 1
Fig. 3.—Images of M33 at (a) H, (b) B band, (c) K band, (d) 24 m, (e) 70 m, and ( f ) 160 m. The H and 24 m data have been convolved to the 70 m
resolution of 1800 . The field of view is approximately 38 0 ; 44 0. North is up, and east is to the left.
Plate 2