1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{emulateapj}
3:
4: \newcommand{\vdag}{(v)^\dagger}
5: \newcommand{\myemail}{hwu@bao.ac.cn}
6:
7:
8: \shorttitle{}
9: %
10: \shortauthors{H.\ Wu, et al.}
11: %\shortauthors{H.\ Wu, C.\ Cao, C.N.\ Hao, F.S.\ Liu, J.L.\ Wang, X.Y.\ Xia, Z.G.\ Deng \& C.K.S.\ Young}
12: \def\comment#1{\bf ???: #1}
13:
14: \received{}
15: \accepted{}
16:
17: \begin{document}
18:
19: \title{
20: PAH AND MID-IR LUMINOSITIES AS MEASURES OF STAR-FORMATION RATE IN
21: {\it SPITZER FIRST LOOK SURVEY} GALAXIES
22: }
23:
24: \author{
25: HONG WU\altaffilmark{1,2},
26: CHEN CAO\altaffilmark{1},
27: CAI-NA HAO\altaffilmark{1},
28: FENG-SHAN LIU\altaffilmark{1},
29: JIAN-LING WANG\altaffilmark{1,3},
30: XIAO-YANG XIA\altaffilmark{3},
31: ZU-GAN DENG\altaffilmark{4},
32: C.\ KE-SHIH YOUNG\altaffilmark{1,5}
33: }
34:
35: \email{hwu@bao.ac.cn}
36: \altaffiltext{1}{National Astronomical Observatories, Chinese Academy of Sciences, Beijing 100012, P.R.\ China}
37: \altaffiltext{2}{Visiting Scholar, Harvard-Smithonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138}
38: \altaffiltext{3}{Department of Physics, Tianjin Normal University, Tianjin 300074, P.R.\ China}
39: \altaffiltext{4}{College of Physical Sciences, Graduate School of the Chinese Academy of Sciences, P.O.\ Box 3908, Beijing 100039, P.R.\ China}
40: \altaffiltext{5}{Department of Physics, University of Hong Kong, Pokfulam Road, Hong Kong, P.R.\ China}
41:
42:
43:
44: \begin{abstract}
45: We present evidence that the 8~$\mu$m (dust) and 24~$\mu$m luminosities of star-forming
46: galaxies are both strongly correlated with their 1.4~GHz and H$\alpha$ luminosities over a
47: range in luminosity of two-to-three orders of magnitude. At the bright end, the correlations
48: are found to be essentially linear over a luminosity range of about two orders of magnitude
49: (corresponding to star-formation rates of several-tenths to several tens of solar masses per
50: year). However, at the faint end, there appears to be a slope change for dwarf
51: galaxies, possibly due to the lower dust-to-gas ratios and lower metallicities of the dwarfs.
52: The correlations suggest that PAH features and mid-IR continuum emissions are good measures
53: of the star formation rates of galaxies, and we present calibrations of star-formation rates
54: based on existing radio and H$\alpha$ relations. Our findings are based on a sample of
55: star-forming galaxies selected from the main field of the Spitzer First Look Survey with the
56: aid of spectroscopic data from the Sloan Digital Sky Survey and VLA 1.4~GHz data.
57: \end{abstract}
58:
59: \keywords { galaxies: starburst -- infrared: galaxies -- stars: formation}
60:
61: \section{INTRODUCTION}
62: \label{sec intro}
63:
64: Star formation rates (SFRs) are fundamental measures of galaxy formation and evolution.
65: Reliable measurement of SFRs could provide us with important clues and constraints
66: concerning the evolution of galaxies. With the less extinction and the less
67: contamination from cirrus, the mid-infrared (MIR) might be ideal tracer of star
68: formation. In addition, broad emission MIR features, such as
69: 3.3, 6.2, 7.7, 8.6, 11.2, 12.7, 16.4, 17.1~$\mu$m, have already been found to be
70: present in most star-forming galaxies by both the Infrared Space Observatory (ISO,
71: Kessler et al.\ 1996; Sturm et al.\ 2000) and Spitzer Space Telescope (Smith et al.\ 2004),
72: and these features are thought to be related to the vibrational emissions of polycyclic
73: aromatic hydrocarbons (PAHs).
74:
75: The ISO 6.75~$\mu$m band just covers the PAH features whilst the 15~$\mu$m band just
76: covers the dust continuum from very small grains (VSGs). However, it turns out that PAH
77: emission from the 12.7~$\mu$m feature can also appear in the 15~$\mu$m band, so observed
78: 15~$\mu$m fluxes will not be due to VSG continua exclusively. Elbaz et al.\ (2002) have
79: demonstrated a relationship between MIR (both 6.75$~\mu$m and 15$~\mu$m) and
80: IR luminosities; whilst Roussel et el.\ (2001) and F\"{o}rster-Schreiber et al.\
81: (2004) have demonstrated one between MIR (both 6.75$~\mu$m and 15$~\mu$m) and
82: H$\alpha$ luminosities in spiral disks.
83:
84: With the launch of the Spitzer Space Telescope (Werner et al.\ 2004), it become available
85: to determine SFRs of millions of galaxies in the MIR bands. The IRAC~8$\mu$m band covers the
86: strongest 7.7~$\mu$m PAH feature, whilst the MIPS 24$\mu$m band avoids most of the PAH
87: features and can sample the continuum emission of VSGs better than the ISO~15$\mu$m band can.
88: Calzetti et al.\ (2005) first obtained the tight relations between Spitzer MIR (8$~\mu$m and
89: 24$~\mu$m) and P$\alpha$ luminosities based on the new Spitzer observations of the 42 knots
90: on M51. The goal of our work is to examine whether Spitzer Space Telescope IRAC 8~$\mu$m
91: and MIPS~24$\mu$m luminosities can be used to measure SFRs of whole galaxies through
92: comparisons with known SFR indicators such as H$\alpha$ and radio luminosities. However,
93: since the MIR continuum can easily be contaminated by emissions from dust heated by AGNs,
94: special care is needed to avoid galaxies with AGNs.
95:
96: In Section 2, we describe how we define our sample of star-forming galaxies.
97: Correlation analysis between MIR luminosities and radio/H$\alpha$ luminosities,
98: and calibrations of MIR luminosities to SFRs are presented in Section 3 and 4.
99: Throughout this paper, we adopt a cosmology with $\Omega_M=0.3$, $\Omega_{\Lambda}=0.7$, and $H_0=70$.
100:
101: %
102: \section{SAMPLE SELECTION}
103: %
104: Our galaxy sample was taken from the extragalactic component of the main field of the Spitzer
105: First Look Survey (FLS).
106: In particular, it was drawn from the overlap area of about 3.7 degree$^2$ of the main field
107: that has been imaged by both instruments: IRAC (3.6, 4.5, 5.8, 8.0~$\mu$m) and MIPS
108: (24$\mu$m). This area has also been covered by the SDSS (Stoughton et al.\ 2002). The IRAC
109: images (in all four IRAC bands) were mosaiced from the ``Basic Calibrated Data'' (BCD)
110: as described by Fazio, et al.\ (2004) and Huang et al.\ (2004); whilst the MIPS images were
111: mosaiced from post-BCD images. Matching the sources detected
112: by SExtractor (Bertin \& Arnouts 1996) in all five bands with the 2mass sources
113: ensured astrometric uncertainties of less than 0.1$\arcsec$. The IRAC and MIPS bands have flux
114: calibration accuracies of better than 10\% (Fazio et al.\ 2004; Rieke et al.\ 2004).
115:
116: FLS sources were then matched with the main galaxy sample of the second data release of the
117: SDSS (Strauss et al.\ 2002) with a matching radius of 2$\arcsec$. The galaxies with
118: H$\alpha$ fluxes of S/N~$\ge$~5 were selected. Based on the traditional line-diagnostic
119: diagram [NII]/H$\alpha$ versus [OIII]/H$\beta$ (Veilleux and Osterbrock 1987; Wu et al.\ 1998),
120: narrow-line AGNs were removed. The fluxes of all of the emission lines were taken from the SDSS
121: catalogue of measured line fluxes (Tremonti et al.\ 2004, version 5.0$\_$4). We also included
122: 1.4~GHz fluxes when the coordinates of VLA 1.4~GHz radio catalogue of Condon et al.\ (2003)
123: matched the sample galaxies to an accuracy of 0.5$~\arcsec$.
124:
125: Our sample of the star-forming galaxies contains a total of 91 galaxies with both H$\alpha$ and MIR
126: fluxes in all four IRAC bands. 70 of these also have MIPS 24$\mu$m fluxes. Due to sensitivity
127: limitations affecting the VLA observations (Condon et al.\ 2003), only 38 of our 91 IRAC sources and
128: 33 of our 70 MIPS sources have 1.4~GHz fluxes. All these objects are local with redshifts of less than 0.2.
129: There are 12 dwarf galaxies with absolute $B$ magnitudes fainter than $-$18.0 in our sample,
130: but only three of these dwarfs have radio data.
131:
132: \section{MIR LUMINOSITIES AND CORRELATION ANALYSIS}
133: \label{radio_ha}
134:
135: To examine whether MIR 8$\mu$m and 24$\mu$m luminosities can be taken as measures of the
136: SFRs of star-forming galaxies, we first compare them with 1.4~GHz and H$\alpha$ luminosities,
137: which have already been extensively used as measures of SFRs.
138: To correct for redshift effects, we adopt a steep non-thermal radio spectrum whose flux obeys a
139: power law of the form $\nu^{-\alpha}$ (Cardiel et al.\ 2003) and has a spectral index of $\alpha$ = 0.8
140: (Condon 1992), and compute rest-frame radio luminosities at 1.4 GHz.
141:
142: As for the H$\alpha$ luminosities, both extinction and aperture corrections were applied to
143: derive the total H$\alpha$ emission of the whole galaxy based on the fibre spectra.
144: The extinction correction included both the foreground extinction of the Milky Way and the
145: intrinsic extinction of galaxy itself. The Milky Way extinction was corrected by the
146: parameterized curve of Cardelli et al.\ (1989) and the intrinsic extinction $A_{H\alpha}$
147: was obtained from the Balmer decrement $F_{H\alpha}/F_{H\beta}$ (Calzetti 2001). For the
148: aperture corrections, we followed the Formula~A1 of Hopkins et al.'s (2003). The corrections
149: were based on the $r$ band fiber magnitude and the corresponding Petrosian magnitude, which can
150: be approximately treated as the total flux of the whole galaxies in $r$ band.
151:
152: Using SExtractor, ``AUTO" magnitudes were obtained for the total luminosities of the sample galaxies
153: in the IRAC and MIPS 24$\mu$m bands. We adopted the SED of the normal HII galaxy NGC~3351 from IRS observation
154: of the Spitzer Legacy Program SINGS (Kennicutt et al.\ 2003) as the template for the K-corrections for
155: our sample galaxies in both the 24$\mu$m and 8$\mu$m bands. To estimate the uncertainty of SED template,
156: we compared it with those of other 20 SINGS normal galaxies, M82, and the models of normal and starburst galaxies of
157: Lagache, Dole and Puget (2003). Within redshift of 0.2, the difference of K-correction is at most 30\%.
158: Although PAH emissions dominate the 8.0$\mu$m band for most of the sample galaxies, there is still
159: a stellar continuum in this band, especially for the galaxies of earlier type. To remove the stellar
160: contribution in this band, we made use of the 3.6$\mu$m band which avoids the strong PAH emission
161: and used it to estimate the stellar contribution in the 8.0$\mu$m band. We selected 74 galaxies
162: without optical emission lines in the FLS and assume they are dust-free at 3.6$\mu$m.
163: The statistics show that the averaged magnitude differences between the 3.6$\mu$m and
164: 8.0$\mu$m bands for these galaxies is about 1.45~mag with a scatter of 0.28~mag for $z < 0.2$.
165: This corresponds the factor of 0.26 to scale the stellar continuum of 3.6$\mu$m to that of 8$\mu$m,
166: and a little larger than the 0.232 of Helou et al.\ (2004). This factor was then used to
167: subtract the stellar continuum contributions from the estimated luminosities for sample galaxies
168: in the 8$\mu$m band. Generally, the stellar continuum
169: flux amounted to about 10\% of the 8.0$\mu$m flux, similar to that of Engelbracht et al.\ (2005).
170: The resulting luminosities which represent the dust emissions of our sample galaxies will henceforth
171: be denoted 8$\mu$m(dust). Though the PAH emissions dominate the 8$\mu$m band, but the dust grains with
172: size $>$ 50$\rm \r{A}$ can still contribute to the 8$\mu$m continuum at some level (Li \& Draine 2001).
173:
174: Figure~\ref{fig1} shows (a) 24$\mu$m and (b) 8$\mu$m(dust) luminosity versus the 1.4 GHz
175: radio luminosity in the log-log space. We can see that a good correlation exists between the MIR and
176: radio luminosities over a luminosity range of about 2 orders of magnitudes. The unfilled symbols
177: represent the only three dwarf galaxies with catalogued radio fluxes. These three dwarfs seem to
178: follow similar correlations to those found for brighter galaxies. The solid lines in the figures are the
179: best fits to all brighter galaxies with two-variable regression. The dotted lines are the linear fits.
180: The fitted parameters and the correlation coefficients are listed in Table~1.
181: As evident from Figure~\ref{fig1}(a), our results are consistent with those of Appleton et al.\ (2004).
182: However, our correlation exhibits a greatly reduced scatter of 0.17 cf.\ Appleton et al.'s (2004) 0.27.
183: In the case of Figure~\ref{fig1} (b), our results appear to be consistent
184: with those of Elbaz et al.\ (2002) at the faint end but there is a departure at the bright end.
185: The correlation was deduced from the Elbaz et al.'s relation of the bolometric IR luminosity versus
186: ISO 6.75$\mu$m luminosity and the radio "$q$" parameter.
187:
188: Figure~\ref{fig2} shows (a) 24$\mu$m and (b) 8$\mu$m(dust) luminosity versus H$\alpha$ luminosity in
189: log-log space. We can observe a good correlation between these two MIR luminosities and the H$\alpha$
190: luminosity for the sample galaxies except that there appears to be a slope change for dwarf galaxies.
191: Both of the linear correlations for non-dwarf galaxies have correlation coefficients of about 0.9.
192: As evident from Figure~\ref{fig2} (b) broad agreement is found with the results of Roussel et al.\ (2001)
193: which were based on a straight-line fit to 6.75$\mu$m (ISO) and H$\alpha$ luminosities for spiral disks.
194:
195: \section{CALIBRATION OF 24$\mu$m AND 8$\mu$m(dust) STAR FORMATION RATES}
196: \label{subsec pah}
197:
198: From the strengths of the correlations listed in Table~1, there appears to be a strong link between both
199: MIR 24$\mu$m and 8$\mu$m(dust) luminosity and both 1.4~GHz radio and H$\alpha$ luminosity.
200: As both radio and H$\alpha$ luminosities are already considered to be good
201: measures of SFR, we can estimate SFRs based on MIR (24$\mu$m and 8$\mu$m(dust)) luminosities
202: from either radio and/or H$\alpha$ SFRs.
203:
204: Combined with the linear relations in Table~1 and the SFR-radio luminosity relation given by Yun, Reddy and
205: Condon (2001), we can derive SFRs from both the 24$\mu$m and 8$\mu$m(dust) luminosities:
206: \begin{equation}
207: \label{sfr24_vla}
208: SFR_{24\mu\,m}(M_{\odot} yr^{-1}) = \frac{\nu\,L_{\nu}[24\mu\,m]}{6.66\times 10^{8} L_{\odot}},
209: \end{equation}
210: \begin{equation}
211: SFR_{8\mu\,m(dust)}(M_{\odot} yr^{-1}) = \frac{\nu\,L_{\nu}[8\mu\,m(dust)]}{1.39\times 10^{9} L_{\odot}}.
212: \end{equation}
213:
214: Similarly, we can also derive SFRs from H$\alpha$ luminosities based on Kennicutt's (1998) SFR formula:
215: \begin{equation}
216: SFR_{24\mu\,m}(M_{\odot} yr^{-1}) = \frac{\nu\,L_{\nu}[24\mu\,m]}{6.43\times 10^{8} L_{\odot}},
217: \end{equation}
218: \begin{equation}
219: \label{sfr8_ha}
220: SFR_{8\mu\,m(dust)}(M_{\odot} yr^{-1}) = \frac{\nu\,L_{\nu}[8\mu\,m(dust)]}{1.57\times 10^{9} L_{\odot}}.
221: \end{equation}
222: However, due to the change in slope exhibited by the dwarf galaxies in Figure~\ref{fig2}, the latter two
223: relations are necessarily limited to the normal star-forming galaxies.
224:
225: There are many factors which could affect calibrations and increase the scatter in the
226: correlations. These include the accuracy of the fibre-aperture corrections, the validity of using the Balmer
227: decrement to estimate obscuration in galaxies (Hopkins et al.\ 2003), possible radio and MIR contamination
228: from obscured weak non-optical AGNs, the template used for K-correction, and so on.
229: We have therefore compared the SFRs derived from
230: formulae (1) through (4). Reassuringly, we found that the agreement between the SFRs based on radio and
231: H$\alpha$ calibrations are around 10\% for both the 24$\mu$m and 8$\mu$m(dust) relations respectively.
232: This level of error is well within the 1-$\sigma$ scatters as listed in table 1.
233:
234: \section{DISCUSSION}
235: \label{sec discuss}
236:
237: Based on the relations we have derived, we suggest that the 24$\mu$m continuum could be due to the
238: warm dust component associated with star formation regions (O'Halloran et al.\ 2005). This support
239: the case made by Dale et al.\ (2001) that the 20-42 $\mu$m continuum may be the best dust emission
240: tracer of current star formation in galaxies. The results of the HII knots in M51 (Calzetti et al.\ 2005)
241: and the 75 nearby galaxies from SINGS (Dale et al.\ 2005) also suggest that the 24$\mu$m emission is
242: a useful tracer of the local star formation. Furthermore, the agreement between our radio-24$\mu$m
243: relation for star-forming galaxies and Appleton et al.'s (2004) relation for radio-quiet sources in the same
244: Spitzer FLS region (Figure~\ref{fig1} (a)), indicates that even for many optical AGNs, the 24$\mu$m emission is
245: still dominated by star formation. The much larger scatter in the radio-24$\mu$m relation for radio-quiet
246: sources could then be explained by the extra contribution from the central AGN.
247:
248: Though we have drawn the linear calibration from MIR luminosity to SFRs, the best fit
249: slopes for the 8$\mu$m(dust) luminosity versus H$\alpha$/radio luminosity is obviously shallower than
250: those for 24$\mu$m (Table~1). Calzetti et al.\ (2005) obtained a similar results for HII knots in M51.
251: Their 24$\mu$m and P$\alpha$ luminosities are nearly linear correlated, but the slope for 8$\mu$m
252: luminosity versus P$\alpha$ luminosity is much shallower than one.
253: Peeters, Spoon and Tielens (2004) have pointed out that PAHs may be better tracers of B stars than
254: tracers of massive star formation (O stars). PAH features can be strong in moderate UV radiation fields
255: such as photo-dissociation regions, but the PAH carriers could be destroyed in hard radiation field
256: (Galliano et al.\ 2005). If it is true, 8$\mu$m(dust) flux might not be a linear measure of intensive
257: starburst activity, this could explain the shallower slope of 8$\mu$m(dust). Alternatively, the VSG continuum
258: might be a useful tracer of star formation in higher UV-radiation fields (F\"{o}rster-Schreiber et al.\ 2004),
259: though VSGs might also be destroyed in very harsh radiation fields (Contursi et al.\ 2000).
260: Since the sample used here are star-forming galaxies, a sample of luminous or even ultraluminous infrared
261: galaxies is needed to check if such relations could be extend to intensive starburst environments.
262:
263: Kewley et al.\ (2002) showed that SFRs derived from H$\alpha$ emission are similar to those derived
264: from IR emission even at SFR of 0.01 $M_{\odot}$ yr$^{-1}$. However Figure~\ref{fig2} shows that the
265: dwarf galaxies systematically deviate from the correlation of the star-forming galaxies to the side of
266: weak PAH emission in the log-log space.
267: Madden (2000) and Dale et al.\ (2005) have also shown that PAH features are weak in dwarf galaxies.
268: It is well known that many dwarf galaxies are gas-rich and have higher gas-to-dust ratios
269: (as well as lower metallicities) than normal galaxies.
270: In low metallicity environments, the UV radiation
271: field is presumably strong enough to destroy the PAH carriers and weaken the PAH features, for example,
272: Blue compact dwarf galaxies SBS~0335-052 (Houck et al.\ 2004). Estimating the oxygen-abundance with
273: the emission line ratios (Kobulnicky, Kennicutt, \& Pizagno 1999),
274: we find that two of four dwarf galaxies (including the most deviant one) exhibit much lower metallicities
275: (about one half or one-third of the solar metallicity) than those of the normal star-forming galaxies.
276: So the slope of the MIR-SFR relations might therefore be metallicity dependent. This is also supported by
277: Engelbracht et al.\ (2005)'s result that the 8$\mu$m-to-24$\mu$m flux ratio strongly depends on metallicity
278: for a sample of 34 star-forming galaxies.
279:
280: \section{ACKNOWLEDGEMENTS}
281: \label{sec acknowledge}
282:
283: The authors would like to thank Jia-sheng Huang and Zhong Wang for useful discussion and advice.
284: We gratefully acknowledge the anonymous referee for insightful comments to improve the paper.
285: This project was supported by the NSF of China through grants No.~10273012, 10333060, 10473013 and
286: NKBRSF~G1999075404; We thanks the Spitzer Science Center operating the Spitzer Archive.
287: We also thank the SDSS team for creation and distribution of the SDSS Archive.
288:
289: \begin{thebibliography}{99}
290:
291: \bibitem[]{356} Appleton, P. N., et al. 2004, \apjs, 154, 147
292: \bibitem[]{357} Bertin, E. \& Arnouts, S. 1996, \aaps, 117, 393
293: \bibitem[]{358} Calzetti, D., et al. 2005, astro-ph/0507427
294: \bibitem[]{359} Calzetti, D. 2001, \pasp, 113, 1449
295: \bibitem[]{360} Cardelli, J. A., Clayton, G. C., \& Mathis, J. S. 1989, \apj, 345, 245
296: \bibitem[]{361} Cardiel, N., Elbaz, D., Schiavon, R. P., Willmer, C. N. A., Koo, D. C., Phillips, A. C.,
297: \& Gallego, J. 2003, \apj, 584, 76
298: \bibitem[]{362} Contursi, A., et al. 2000, \aap, 362, 310
299: \bibitem[]{363} Condon, J. J., et al. 2003, \aj, 125, 2411
300: \bibitem[]{364} Condon, J. J. 1992, \araa, 30, 575
301: \bibitem[]{365} Dale, D. A., et al. 2005, astro-ph/0507645
302: \bibitem[]{366} Dale, D. A., Helou, G., Contursi, A., Silbermann, N. A., \& Kolhatkar, S. 2001, \apj, 549, 215
303: \bibitem[]{367} Engelbracht, C. W., Gordon, K. D., Rieke, G. H., Werner, M. W., Dale, D. A., \& Latter, W. B. 2005, \apj, 628, L29
304: \bibitem[]{368} Elbaz, D., Cesarsky, C. J., Chanial, P., Aussel, H., Franceschini, A., Fadda, D., \& Chary, R. R. 2002, \aap, 384, 848
305: \bibitem[]{369} Fazio, G. G., et al. 2004, \apjs, 154, 10
306: \bibitem[]{370} F\"{o}rster-Schreiber, N. M., Roussel, H., Sauvage, M., \& Charmandaris, V. 2004, \aap, 419, 501
307: \bibitem[]{371} Galliano, F., Madden, S. C., Jones, A. P., Wilson, C. D., \& Bernard, J.-P. 2005, astro-ph/0501632
308: \bibitem[]{372} Helou, G., et al. 2004, \apjs, 154, 253
309: \bibitem[]{373} Hopkins, A. M., et al. 2003, \apj, 599, 971
310: \bibitem[]{374} Houck, J. R., et al. 2004, \apjs, 154, 211
311: \bibitem[]{375} Huang, J. S., et al. 2004, \apjs, 154, 44
312: \bibitem[]{376} Kennicutt, R. C., Jr., et al. 2003, \pasp, 115, 928
313: \bibitem[]{377} Kennicutt, R. C., Jr. 1998, \araa, 36, 189
314: \bibitem[]{378} Kessler, M. F., et al. 1996, \aap, 315, L27
315: \bibitem[]{379} Kewley, L. J., Geller, M. J., Jansen, R. A., \& Dopita, M. A. 2002, \aj, 124, 3135
316: \bibitem[]{380} Kobulnicky, H. A., Kennicutt, R. C. Jr, \& Pizagno, J. L. 1999, \apj, 514, 544
317: \bibitem[]{381} Li, A., \& Draine, B. T. 2001, \apj, 554, 778
318: \bibitem[]{382} Lagache, G., Dole, H., \& Puget, J. L. 2003, \mnras, 338, 555
319: \bibitem[]{383} Madden, S. C. 2000, New Astronomy Review, 44, 249
320: \bibitem[]{384} O'Halloran, B., McBreen, B., Metcalfe, L., Delaney, M., \& Coia, D. 2005, astro-ph/0506037
321: \bibitem[]{385} Peeters, E., Spoon, H. W. W., \& Tielens, A. G. G. M. 2004, \apj, 613, 986
322: \bibitem[]{386} Rieke, G. H., et al. 2004, \apjs, 154, 25
323: \bibitem[]{387} Roussel, H., Sauvage, M., Vigroux, L., \& Bosma, A. 2001, \aap, 372, 427
324: \bibitem[]{388} Smith, J. D. T., et al. 2004, \apjs, 154, 199
325: \bibitem[]{389} Stoughton, C., et al. 2002, \aj, 123, 485
326: \bibitem[]{390} Strauss, M. A., et al. 2002, \aj, 124, 1810
327: \bibitem[]{391} Sturm, E., et al. 2000, \aap, 358, 481
328: \bibitem[]{392} Tremonti, C. A. et al. 2004, \apj, 613, 898
329: \bibitem[]{393} Veilleux, S., \& Osterbrock, D. E. 1987, \apjs, 63, 295
330: \bibitem[]{394} Werner, M. W., et al. 2004, \apjs, 154, 1
331: \bibitem[]{395} Wu, H., Zou, Z. L., Xia, X. Y., \& Deng, Z. G. 1998, \aaps,132, 181
332: \bibitem[]{396} Yun, M. S., Reddy, N. A., \& Condon, J. J. 2001, \apj, 554, 803
333:
334: \end{thebibliography}
335:
336: \clearpage
337:
338: \begin{figure}
339: \centerline{\includegraphics[height=0.9\textwidth]{f1.eps}}
340: \caption{The correlations between the MIR and radio luminosities. The filled circles are
341: the sample galaxies and the unfilled symbols are the dwarf galaxies with $M_B$ $< -18.0$~mag.
342: The solid and dotted lines are the best non-linear and linear fits respectively for
343: star-forming galaxies. (a) 24$\mu$m versus radio (1.4~GHz). The long-dashed line represents
344: the linear fit to the radio and 24$\mu$m luminosities obtained by Appleton et al.\ (2004)
345: for radio-quiet sources with K-corrections based on an M82 template. (b) 8$\mu$m(dust)
346: versus radio. The short-dashed line represents the relation of the radio and 6.75$\mu$m (ISO)
347: luminosities drawn from Elbaz et al.\ (2002).
348: } \label{fig1}
349: \end{figure}
350:
351: \clearpage
352:
353: \begin{figure}
354: \centerline{\includegraphics[height=0.9\textwidth]{f2.eps}}
355: \caption{The correlations between the MIR and H$\alpha$ luminosities.
356: The filled, unfilled circles, the solid and dotted lines are as in Figure~1.
357: (a) 24$\mu$m versus H$\alpha$. (b) 8$\mu$m(dust) versus H$\alpha$. The dot-dashed line is
358: the linear relation between ISO 6.75$\mu$m and H$\alpha$ luminosities for spiral disks found
359: by Roussel et al.s (2001).
360: } \label{fig2}
361: \end{figure}
362:
363: \clearpage
364:
365: \begin{table}
366: \begin{center}
367: \caption{Correlation coefficients for luminosity relations for star-forming galaxies}
368: \label{tab correlation}
369: \small
370: \begin{tabular}{llrrrrrr}
371: \tableline\tableline
372: $y$ & $x$ & a & b & r & c & N & \\
373: \tableline
374: $\nu L_{\nu}[24\mu m](L_{\odot})$ &$ L_{1.4GHz}(W\ Hz^{-1})$ &-13.31$\pm$0.42&1.04$\pm$0.09& 0.82&-12.41$\pm$0.17& 30 \\
375: $\nu L_{\nu}[8\mu m](L_{\odot})$ &$ L_{1.4GHz}(W\ Hz^{-1})$ &-10.12$\pm$0.31&0.91$\pm$0.07& 0.88&-12.09$\pm$0.13& 35 \\
376: $\nu L_{\nu}[24\mu m](L_{\odot})$ &$ L_{H\alpha}(L_{\odot})$& 0.34$\pm$0.19&1.12$\pm$0.07& 0.88& 1.29$\pm$0.19& 63 \\
377: $\nu L_{\nu}[8\mu m](L_{\odot})$ &$ L_{H\alpha}(L_{\odot})$ & 2.31$\pm$0.15&0.92$\pm$0.05& 0.87& 1.68$\pm$0.19& 79\\
378: \tableline
379: \end{tabular}\\
380: \end{center}
381: {a and b are the coefficients of the non-linear fitting of $\log_{10}(y)=a+b\log_{10}(x)$ and c is the
382: coefficient of the linear fitting of $\log_{10}(y)=c+\log_{10}(x)$. $r$ is the Spearman's correlation
383: coefficient, rho, and $N$ is the number of sample galaxies (excluding the dwarf galaxies) used for the
384: fitting.}
385: \end{table}
386:
387:
388:
389:
390:
391:
392:
393:
394:
395:
396:
397:
398:
399:
400:
401:
402:
403:
404: \label{lastpage}
405: \end{document}
406: