0711.4363/ms.tex
1: %% Referee manuscript produces a one-column, double-spaced document:
2: %\documentclass[12pt,manuscript]{aastex}
3: 
4: %% preprint produces a one-column, single-spaced document:
5: %\documentclass[12pt,preprint]{aastex}
6: % preprint2 produces a double-column, single-spaced document:
7: %\documentclass[preprint2]{aastex}
8: 
9: %% Emulate ApJ document - for a nice 2 column compact form
10: \documentclass{emulateapj}
11: 
12: 
13: \usepackage{amssymb}
14: \usepackage{graphicx}
15: 
16: 
17: \slugcomment{Accepted by ApJ, November 27th, 2007}
18: 
19: 
20: 
21: \newcommand{\degree}{\ensuremath{^\circ}}
22: 
23: \begin{document}
24: 
25: 
26: %% LaTeX will automatically break titles if they run longer than
27: %% one line. However, you may use \\ to force a line break if
28: %% you desire.
29: 
30: \title{Blue Compact Dwarf Galaxies with Spitzer: The Infrared/Radio Properties}
31: 
32: \author{Yanling Wu\altaffilmark{1}, V. Charmandaris\altaffilmark{2,3}, J.~R. Houck\altaffilmark{1}, 
33:   J. Bernard-Salas\altaffilmark{1},V.  Lebouteiller\altaffilmark{1}, B.~R. Brandl\altaffilmark{4}, 
34:   D. Farrah\altaffilmark{1}}
35: 
36: \altaffiltext{1}{Astronomy Department, Cornell University, Ithaca, NY
37:   14853}
38: 
39: \altaffiltext{2}{University of Crete, Department of Physics, GR-71003,
40: Heraklion, Greece}
41: 
42: \altaffiltext{3}{IESL/Foundation for Research and Technology - Hellas,
43:   GR-71110, Heraklion, Greece and Chercheur Associ\'e, Observatoire de
44:   Paris, F-75014, Paris, France}
45: 
46: \altaffiltext{4}{Leiden Observatory, Leiden University, P.O. Box 9513,
47: 2300 RA Leiden, The Netherlands}
48: 
49: \email{wyl@astro.cornell.edu, vassilis@physics.uoc.gr,jrh13@cornell.edu,
50:   jbs@isc.astro.cornell.edu,vianney@isc.astro.cornell.edu, 
51:   brandl@strw.leidenuniv.nl,duncan@isc.astro.cornell.edu}
52: 
53: 
54: \begin{abstract}
55:   We study the correlation between the radio, mid-infrared and
56:   far-infrared properties for a sample of 28 blue compact dwarf (BCD)
57:   and low metallicity star-forming galaxies observed by {\em Spitzer}.
58:   We find that these sources extend the same far-infrared to radio
59:   correlation typical of star forming late type galaxies to lower
60:   luminosities. In BCDs, the 24$\mu$m (or 22$\mu$m) mid-infrared to
61:   radio correlation is similar to starburst galaxies, though there is
62:   somewhat larger dispersion in their q$_{24}$ parameter compared to
63:   their q$_{\rm FIR}$. No strong correlations between the q parameter
64:   and galaxy metallicity or effective dust temperature have been
65:   detected, though there is a hint of decreasing q$_{24}$ at low
66:   metallicities. The two lowest metallicity dwarfs in our sample,
67:   IZw18 and SBS0335-052E, despite their similar chemical abundance,
68:   deviate from the average q$_{24}$ ratio in opposite manners,
69:   displaying an apparent radio excess and dust excess respectively.
70: 
71: \end{abstract}
72: 
73: \keywords{
74:   galaxies:starburst ---
75:   galaxies: dwarf ---
76:   infrared: galaxies ---
77:   radio continuum: galaxies
78:   }
79: 
80: \section{Introduction}
81: 
82: More than 35 years ago, a correlation between the 10$\mu$m
83: mid-infrared (mid-IR) and 1.4\,GHz luminosities in galactic nuclei was
84: first noticed by \citet{van71}. In 1983, the launch of the {\em
85:   Infrared Astronomical Satellite (IRAS)} brought a new era in
86: collecting large samples of data in the infrared, and sparked numerous
87: studies investigating the above mentioned correlation in star forming
88: galaxies. Both the radio and infrared emission are closely
89: related to star formation activities. The radio continuum is known to
90: originate from two processes: thermal free-free emission from ionized
91: gas in H{\sc ii} regions and non-thermal synchrotron radiation from
92: relativistic electrons that are accelerated in the supernovae remnants
93: (SNRs) \citep[see][ for a review]{Condon92}. The thermal emission
94: usually has a rather flat spectrum with
95: f$_{\nu}$\,$\propto$\,$\nu^{-0.1}$ while the non-thermal component
96: often has a much steeper spectral slope with
97: f$_{\nu}$\,$\propto$\,$\nu^{-0.8}$.  In normal galaxies the relative
98: contribution of the two components varies with frequency and at
99: 1.4\,GHz the radio continuum is dominated (at nearly $\sim$90\%) by
100: the non-thermal component. The infrared emission is due to the thermal
101: re-radiation of starlight from dust surrounding H{\sc ii} regions
102: \citep{Harwit75}.  Based mostly on {\em IRAS} data, it has been shown
103: that the far-infrared (FIR, 40-120\,$\mu$m) to radio correlation holds
104: well over a remarkably wide range of star forming galaxies, spanning
105: several orders of magnitude in luminosities \citep{Helou85, Dejong85,
106:   Condon91,Yun01}.  The availability of deep observations of distant
107: galaxies by the {\em Infrared Space Observatory (ISO)} and recently by
108: {\em Spitzer} \citep{Werner04} showed that the correlation between the
109: IR and radio emission is not limited to the local universe but also
110: extends in galaxies at higher redshifts \citep{Garrett02,Appleton04}.
111: It was also revealed that in addition to the FIR, the mid-IR emission
112: at $\sim$24$\mu$m correlates with the 20cm radio continuum, though
113: with more scatter \citep{Elbaz02,Gruppioni03,Appleton04,Wu05}. A
114: number of very deep {\em Spitzer} mid-IR and FIR surveys can now probe
115: a population of galaxies with low infrared luminosities for which
116: ancillary data, including deep radio imaging, are becoming available
117: \citep[e.g.][]{Jannuzi99,Sanders07,Rosenberg06}. This is particularly
118: interesting since in the near future $\sim20\mu$m is the longest
119: wavelength which will likely be probed by the James Web Space
120: Telescope.  It is thus instructive to examine the mid-IR to radio
121: correlation in more detail in these low luminosity nearby systems, for
122: which little is known to date.
123: 
124: As ubiquitous as the FIR/radio correlation appears to be, there are a
125: few significant deviations from it.  Radio excess exists in
126: some instances, such as galaxies hosting an active galactic nucleus
127: (AGN). External magnetic field compression due to the interaction with
128: nearby galaxies could also produce extra emission in the radio
129: continuum \citep{Miller01}. Conversely, synchrotron deficiency has been
130: found in some nascent starburst galaxies studied by
131: \citet{Roussel03,Roussel06} which was attributed to the lack of time
132: for massive young stars to evolve into supernovae (SN) since these
133: galaxies are just at the onset of a starburst episode. On the other
134: hand, in the infrared, dust emission can be damped in an
135: optically-thin environment, because the ultraviolet (UV) and optical
136: light may not be fully reprocessed by the dust, as seen in low
137: luminosity dwarf galaxies.
138: 
139: 
140: Blue compact dwarf galaxies are a group of extra-galactic objects that
141: are characterized by their blue optical colors, small sizes
142: ($\le$1\,kpc) and low luminosities (M$_B>$-18). These galaxies do not
143: display any AGN signature\footnote{A possible exception among our
144:   sources is CG0752 in which the high ionization lines
145:   [NeV]$\lambda$14.3/24.3$\mu$m are detected \citep[see][]{Hao07}.} and
146: have recent bursts of star formation in a relatively unevolved
147: chemical environment. As such they have been proposed as nearby
148: analogs of star formation in young galaxies in the early universe.  In
149: a metal poor environment, star forming regions are usually optically
150: thin and emit less in the infrared.  However, \citet{Devereux89}
151: suggested that in a low luminosity galaxy, the radio emission also
152: decreases, and probably faster than the infrared.  The deficiency in
153: both the non-thermal radio and FIR emission may counterbalance each
154: other and result in a similar FIR/radio ratio to the one observed in
155: normal spiral galaxies \citep{Klein91}.  This was examined by a study
156: of star formation rates (SFRs) in BCDs performed by \citet{Hopkins02},
157: in which the authors found an excellent agreement between the SFRs
158: estimated from 1.4\,GHz and 60\,$\mu$m luminosities.  As
159: \citet{Bell03} has pointed out though, the FIR/radio correlation is
160: almost linear, not because the IR and radio emission reflect the SFRs
161: correctly, but because in low luminosity galaxies they are both
162: underestimated by similar factors.  This is in agreement with
163: \citet{Helou93}, who found from their modelling work in disk galaxies
164: that the transparency of the disk was about the same to both
165: re-emission processes.  \citet{Hunt05a} in their analysis of the
166: spectral energy distributions (SEDs) of low metallicity BCDs, noticed
167: that these systems do not follow several of the usual correlations
168: between the mid-IR, FIR and radio emission and display a scatter of a
169: factor of $\sim$10.
170: 
171: The {\em Spitzer} Space telescope has enabled us to
172: study the infrared properties of a large sample of BCDs, probing the
173: lower end of the luminosity and metallicity range. In this paper, the
174: fifth in a series \citep{Houck04b,Wu06,Wu07a,Wu07b}, we examine their
175: mid-IR and FIR to radio correlation extending the work of
176: \citet{Hopkins02}. We describe the sample selection and the observational
177: data in \S 2. A detailed study of mid-IR and FIR/radio correlation, as
178: well as its dependence with other parameters, such as metallicity and
179: dust temperature are presented in \S 3.  We also discuss two extreme
180: cases, IZw18 and SBS0335-052 in \S 3. We summarize our conclusions
181: in \S 4.
182: 
183: 
184: \section{Observations}
185: 
186: As part of the IRS\footnote{The IRS was a collaborative venture
187:   between Cornell University and Ball Aerospace Corporation funded by
188:   NASA through the Jet Propulsion Laboratory and the Ames Research
189:   Center.} \citep{Houck04a} Guaranteed Time Observation (GTO) program
190: (PID: 85), we have compiled a sample of BCD candidates ($\sim$64)
191: selected from the Second Byurakan Survey (SBS), Bootes void galaxies
192: \citep{Kirshner81,Popescu00}, and other commonly studied BCDs. These
193: sources are known to have low metallicities ranging from
194: 0.03\,Z$_\odot$ to 0.5\,Z$_\odot$\footnote{Here we adopt the new
195:   oxygen solar abundance of 12+log(O/H)=8.69 \citep{Prieto01}.
196:   \citet{Wu07b} have derived neon and sulfur abundances for a sample
197:   of BCDs using the infrared data, but these abundances are not
198:   available for all of the sources in this study.}. Their 22\,$\mu$m
199: fluxes have been published in \citet{Wu06}. We also include 10
200: galaxies from \citet{Engelbracht05} (PID: 59), which are mostly BCDs
201: and starburst galaxies, and span a larger metallicity range
202: (0.03\,Z$_\odot$$\sim$1.5\,Z$_\odot$)\footnote{The galaxies in PID 59
203:   partly overlap with the BCDs in PID 85. We have also excluded some
204:   galaxies which the authors derived their 24\,$\mu$m fluxes by doing
205:   a color correction to the {\em IRAS} 25\,$\mu$m fluxes because of
206:   the aperture difference between these two band filters.}. For all
207: galaxies with 22\,(24)\,$\mu$m detections, we searched the literature
208: as well as the public archives (NVSS and FIRST) for 1.4GHz radio
209: continuum data. We restirct our sample to sources with both mid-IR and
210: radio detections which results in a sample of 23 galaxies. Finally, we
211: also include 5 galaxies that have 22\,$\mu$m detections and 1.4\,GHz
212: upper limits published by Hopkins et al. (2002) for comparison between
213: ours and Hopkins' samples. Note that the sample was not selected based
214: on infrared properties, but merely on BCD-type objects and the
215: availability of both infrared and radio data.  As a result, our sample
216: is not complete, but the large number of sources that only became
217: detectable in the infrared with {\em Spitzer}, allows us to probe the
218: properties of low luminosity dwarf galaxies, and provide statistically
219: meaningful results. The observational information for this sample and
220: previously published data are presented in Table \ref{tab1}, which
221: includes the positions of the sources, their mid-IR, FIR and 1.4GHz
222: flux densities, as well as oxygen abundances of the ionized gas.
223: 
224: All sources in this study have mid-IR flux measurements either at
225: 22$\mu$m with the IRS red peak-up camera and/or at 24\,$\mu$m with
226: MIPS \citep{Rieke04}. The photometric fluxes of these two bands differ
227: by less than 10\% for galaxies in the local universe.  This was
228: confirmed by using a suite of $\sim$100 spectra of nearby
229: galaxies from our IRS/GTO database and calculating their synthetic 22
230: and 24$\mu$m fluxes after convolving the spectra with the
231: corresponding filter response curves. For consistency, in our analysis
232: we use the MIPS 24\,$\mu$m measurements, and the 22\,$\mu$m values are
233: only used when the 24\,$\mu$m values are not available. For sources
234: above the IRS 22$\mu$m and MIPS 24$\mu$m saturation limits we use the
235: IRS low resolution spectrum to estimate a synthetic 24\,$\mu$m flux.
236: We also obtained far-infrared fluxes for our sample from the archival
237: {\em IRAS} 60 and 100\,$\mu$m data \citep{Moshir90, Sanders03}.
238: 
239: Most of the 1.4\,GHz radio continuum data are from the NRAO VLA Sky
240: Survey (NVSS) \citep{Condon98}, while one is from the Faint Images of
241: the Radio Sky at Twenty cm (FIRST) \citep{Becker95}, along with some
242: individual observations (references in Table \ref{tab1}). A total of
243: seven sources were too faint and were not included in the NVSS
244: catalogue. For those we used the values of \citet{Hopkins02} who
245: studied a similar BCD sample and remeasured the 1.4GHz fluxes using
246: the images from NVSS and FIRST, providing radio detections for 2
247: sources and better upper limits for another five sources that overlaps
248: with the galaxies in our sample.
249: 
250: \begin{deluxetable*}{llrrrrrrccccc}
251:   \tabletypesize{\scriptsize}
252:   \setlength{\tabcolsep}{0.02in}
253:   \tablecaption{Properties of the Sample\label{tab1}}
254:   \tablewidth{0pc}
255:   \tablehead{
256:     \colhead{ID} & \colhead{Object Name} & \colhead{RA} & \colhead{Dec} & \multicolumn{4}{c}{Flux (mJy)} & 
257:     \colhead{12+log(O/H)} & \multicolumn{4}{c}{References} \\
258:     \cline{5-8}\cline{10-13}
259:     \colhead{} & \colhead{} & \colhead{J2000} & \colhead{J2000} & \colhead{1.4\,GHz} & \colhead{24(22)\,$\mu$m} & 
260:     \colhead{60\,$\mu$m} & \colhead{100\,$\mu$m} & \colhead{} & \colhead{1.4\,GHz} & \colhead{24\,$\mu$m} &
261:     \colhead{IRAS} & \colhead{Z} \\
262:   }
263:   \startdata
264:   
265: 1  &  Haro11              & 00h36m52.5s  & -33d33m19s &  26.8  &   1900   &   6880 &   5040 &  7.9  & (1) & (2) & (3) & (4) \\
266: 2  &  NGC1140             & 02h54m33.6s  & -10d01m40s &  23.6  &    316.6 &   3358 &   4922 &  8.5  & (1) & (5) & (6) & (7) \\
267: 3  &  SBS0335-052E        & 03h37m44.0s  & -05d02m40s &  0.46  &     66   & \nodata& \nodata&  7.3  & (8) & (2) &     & (9) \\
268: 4  &  NGC1569             & 04h30m47.0s  & +64d50m59s & 336.3  &   2991   &  54360 &  55290 &  8.2  & (1) & (2) & (3) &(10) \\
269: 5  &  IIZw40              & 05h55m42.6s  & +03d23m32s &  32.5  &   1500   &   6570 &   5270 &  8.1  & (1) & (2) & (3) & (7) \\
270: 6  &  UGC4274             & 08h13m13.0s  & +45d59m39s &  10.5  &    240   &   3236 &   6448 &  8.5  & (1) & (2) & (6) &     \\ 
271: 7  &  He2-10              & 08h36m15.2s  & -26d24m34s &  83.8  &   4900   & \nodata& \nodata&  8.9  & (1) & (2) &     &(11) \\
272: 8  &  NGC2782             & 09h14m05.1s  & +40d06m49s & 124.5  &    960   &   9170 &  13760 &  8.8  & (1) & (2) & (3) &(12) \\
273: 9  &  NGC2903             & 09h32m10.1s  & +21d30m03s & 444.5  &   2200   &  60540 & 130430 &  9.3  & (1) & (2) & (3) &(11) \\
274: 10 &  IZw18               & 09h34m02.0s  & +55d14m28s &   1.83 &      5.5 & \nodata& \nodata&  7.2  &(13) & (2) &     & (9) \\
275: 11 &  SBS0940+544         & 09h44m16.7s  & +54d11m33s &$<$2.3  &      2.3 & \nodata& \nodata&  7.5  &(14) & (5) &     &(15) \\ 
276: 12 &  NGC3077             & 10h03m19.1s  & +68d44m02s &  29.0  &   1500   &  15900 &  26530 &  8.6  & (1) & (2) & (3) &(16) \\
277: 13 &  Mrk153              & 10h49m05.0s  & +52d20m08s &   4.0  &     29   &    280 & $<$480 &  7.8  & (1) & (2) & (6) &(17) \\
278: 14 &  VIIZw403            & 11h27m59.9s  & +78d59m39s &   1.2  &     28   & \nodata& \nodata&  7.7  &(18) & (2) &     & (9) \\
279: 15 &  Mrk1450             & 11h38m35.6s  & +57d52m27s &$<$2.0  &     48   &    279 & $<$575 &  8.0  &(14) & (2) & (6) & (7) \\
280: 16 &  UM448               & 11h42m12.4s  & +00d20m03s &  32.6  &    560   &   4139 &   4321 &  8.0  & (1) & (2) & (6) & (7) \\
281: 17 &  UM461               & 11h51m33.3s  & -02d22m22s &$<$2.6  &     30   & \nodata& \nodata&  7.8  &(14) & (2) &     & (9) \\
282: 18 &  UM462               & 11h52m37.3s  & -02d28m10s &   5.8  &    110   &    944 &    896 &  8.0  & (1) & (2) & (6) & (9) \\
283: 19 &  SBS1159+545         & 12h02m02.4s  & +54d15m50s &$<$2.3  &      6.4 & \nodata& \nodata&  7.5  &(14) & (5) &     &(19) \\ 
284: 20 &  NGC4194             & 12h14m09.5s  & +54d31m37s & 100.7  &   3100   &  23200 &  25160 &  8.8  & (1) & (2) & (3) &(16) \\
285: 21 &  NGC4670             & 12h45m17.1s  & +27d07m32s &  13.7  &    210   &   2634 &   4470 &  8.2  & (1) & (2) & (6) &(12) \\
286: 22 &  SBS1415+437         & 14h17m01.4s  & +43d30m05s &   4.3  &     19.6 & \nodata& \nodata&  7.6  &(14) & (5) &     & (9) \\
287: 23 &  Mrk475              & 14h39m05.4s  & +36d48m21s &$<$2.7  &     10.8 & \nodata& \nodata&  7.9  &(14) & (5) &     & (9) \\
288: 24 &  CG0563              & 14h52m05.7s  & +38d10m59s &   6.2  &     97.0 &    870 &   1900 &  8.7  & (1) &     & (6) &     \\
289: 25 &  CG0752              & 15h31m21.3s  & +47d01m24s &   4.8  &    138.9 &    837 &   1059 &\nodata& (1) & (5) & (6) &     \\
290: 26 &  SBS1533+574         & 15h34m13.8s  & +57d17m06s &   4.2  &     53.4 &    257 &    405 &  8.1  &(14) & (5) & (6) & (9) \\
291: 27 &  Mrk1499             & 16h35m21.1s  & +52d12m53s &   1.5  &     33.8 &    256 &    617 &  8.1  &(20) & (5) & (6) &(21) \\
292: 28 &  Mrk930              & 23h31m58.3s  & +28d56m50s &  12.2  &    170   &   1245 &$<$2154 &  8.1  & (1) & (2) & (6) & (7) \\
293: 
294: \enddata
295: 
296: \tablerefs{(1) \citet{Condon98}, (2) \citet{Engelbracht05}, (3) \citet{Sanders03}, (4) \citet{Bergvall02}, 
297:   (5) \citet{Wu06}, (6) \citet{Moshir90}, (7) \citet{Guseva00}, (8) \citet{Hunt04}, (9) \citet{Izotov99},
298:   (10) \citet{Kobulnicky97}, (11) \citet{Kobulnicky99}, (12) \citet{Heckman98}, (13) \citet{Hunt05a}, 
299:   (14) \citet{Hopkins02}, (15) \citet{Thuan05}, (16) \citet{Storchi94}, (17) \citet{Kunth85}, (18) \citet{Leroy05}, 
300:   (19) \citet{Izotov98}, (20) FIRST catalog, (21) \citet{Shi05}}
301: \end{deluxetable*}
302: 
303: 
304: 
305: Our final sample consists of 28 galaxies, all of which have {\em Spitzer}
306: mid-IR 24\,$\mu$m and/or 22\,$\mu$m flux measurements.  Among these
307: galaxies, 23 sources have 1.4GHz radio continuum data and 5 have
308: measured upper limits. {\em IRAS} 60\,$\mu$m and 100\,$\mu$m fluxes
309: are available for 16 sources and 3 more are detected only at
310: 60\,$\mu$m. We list the photometry of our sources in Table \ref{tab1}.
311: The uncertainty in the 22\,$\mu$m or 24\,$\mu$m photometry is less
312: than 5\%.  The {\em IRAS} 60\,$\mu$m and 100\,$\mu$m fluxes typically
313: have less than a few percent of error, but could go up to $\sim$15\%
314: for some of the fainter sources.  The rms noise level for NVSS is
315: $\sim$0.5\,mJy\,beam$^{-1}$.
316: 
317: 
318: \section{Results}
319: 
320: \subsection{Mid-IR and FIR to Radio Correlation in BCDs}
321: 
322: The sensitivity and efficiency of the {\em Spitzer} Space Telescope
323: has allowed us to probe the correlation between the mid-IR and radio
324: luminosities for a large number of galaxies. Using the 24\,$\mu$m and
325: 70\,$\mu$m MIPS imaging data of the First Look Survey,
326: \citet{Appleton04} have demonstrated the first direct evidence for the
327: universality of the mid-IR/radio and FIR/radio correlation to
328: z$\sim$1. \citet{Wu05} have also studied the mid-IR/radio correlation
329: in a sample of star-forming galaxies and found that both the 8\,$\mu$m
330: and 24\,$\mu$m luminosity are clearly correlated with the 1.4GHz radio
331: luminosity. Their sample included only a few (3) dwarf galaxies and
332: suggested that there may be a slope change for dwarf galaxies,
333: which could be due to the lower dust-to-gas ratios and lower metallicities
334: of the dwarfs.  A detailed analysis of the spatial distribution of the
335: infrared to radio correlation using {\em Spitzer} data on a sample of nearby
336: late type spiral galaxies has been performed by
337: \citet{Murphy06a,Murphy06b} in which the authors found that the ratio
338: of the mid-IR and FIR to radio emission does vary from one region to
339: the other. In general though, the dispersion is small and the ratio
340: decreases with surface brightness and galactocentric radius.
341: 
342: \begin{figure}
343:   \epsscale{1.2}
344:   \plotone{f1.eps}
345:   \caption{The FIR to 1.4\,GHz radio luminosity correlation for this
346:     sample. Our data are presented with filled circles. The numbers
347:     next to each symbol correspond to their IDs in Table \ref{tab1}.
348:     The solid line is the best fit to this sample excluding the upper
349:     limits.  For comparison, we have also include the starburst
350:     galaxies of \citet{Brandl06}, indicated with triangles.}
351:   \label{fig:fig1}
352: \end{figure}
353:  
354: 
355: \begin{deluxetable*}{llrcccrr}
356:   \tabletypesize{\scriptsize}
357:   \setlength{\tabcolsep}{0.02in}
358:   \tablecaption{Derived Quantities of the Sample\label{tab2}}
359:   \tablewidth{0pc}
360:   \tablehead{
361:     \colhead{ID} & \colhead{Object Name} & \colhead{Distance\tablenotemark{a}} & \colhead{L$_{\rm 1.4GHz}$} 
362:     & \colhead{L$_{24\mu m}$} & \colhead{L$_{\rm FIR}$} & \colhead{q$_{24}$} & \colhead{q$_{\rm FIR}$} \\
363:     \colhead{} & \colhead{} & \colhead{(Mpc)} & \colhead{($\times$10$^{20}$\,W\,Hz$^{-1}$)} &
364:     \colhead{($\times$10$^8$L$_\odot$)} & \colhead{($\times$10$^8$L$_\odot$)} & \colhead{} & \colhead{}
365:   }
366:   \startdata
367:   
368: 1  &  Haro11              &  88   & 250.5  & 584.2  &   706.3  & 1.85  & 2.46  \\
369: 2  &  NGC1140             &  21.2 & 12.8   & 5.6    &   24.4   & 1.13  & 2.29  \\
370: 3  &  SBS0335-052E        &  58   & 18.7   & 8.8    &  \nodata & 2.16  &\nodata\\
371: 4  &  NGC1569             &  2.0  & 1.6    & 0.5    &   3.1    & 0.95  & 2.29  \\
372: 5  &  IIZw40              &  12.4 & 6.0    & 9.2    &   13.7   & 1.66  & 2.36  \\
373: 6  &  UGC4274             &  6.9  & 0.6    & 0.5    &   2.8    & 1.36  & 2.68  \\ 
374: 7  &  He2-10              &  12   & 15.6   & 29.9   &  \nodata & 1.77  &\nodata\\
375: 8  &  NGC2782             &  41.7 & 216.3  & 66.3   &   260.4  & 0.89  & 2.00  \\
376: 9  &  NGC2903             &  8.9  & 42.5   & 6.9    &   90.9   & 0.69  & 2.34  \\
377: 10 &  IZw18               &  18.2 & 0.7    & 0.07   &  \nodata & 0.48  & 1.31\tablenotemark{b}\\
378: 11 &  SBS0940+544         &  23   &$<$ 1.5 & 0.05   &  \nodata &$>$0.00&\nodata\\ 
379: 12 &  NGC3077             &  3.8  & 0.5    & 0.9    &   3.9    & 1.71  & 2.89  \\
380: 13 &  Mrk153              &  40.5 & 7.9    & 1.9    &  $<$7.9  & 0.86  &$<$2.00\\
381: 14 &  VIIZw403            &  4.3  & 0.03   & 0.02   &  \nodata & 1.37  &\nodata\\
382: 15 &  Mrk1450             &  20.0 &$<$1.0  & 0.8    &  $<$2.1  &$>$1.38&\nodata \\
383: 16 &  UM448               &  87.4 & 300.6  & 188.9  &   458.6  & 1.23  & 2.19  \\
384: 17 &  UM461               &  13.4 &$<$0.6  & 0.2    &  \nodata &$>$1.06&\nodata\\
385: 18 &  UM462               &  13.4 & 1.3    & 0.8    &   2.4    & 1.28  & 2.29  \\
386: 19 &  SBS1159+545         &  50   &$<$7.0  & 0.6    &  \nodata &$>$0.44&\nodata\\ 
387: 20 &  NGC4194             &  41.5 & 209.3  & 212.0  &   586.0  & 1.49  & 2.45  \\
388: 21 &  NGC4670             &  23.2 & 8.9    & 4.5    &   24.3   & 1.19  & 2.44  \\
389: 22 &  SBS1415+437         &  8.7  & 0.4    & 0.06   &  \nodata & 0.66  &\nodata\\
390: 23 &  Mrk475              &  11.2 &$<$0.4  & 0.05   &  \nodata &$>$0.60&\nodata\\
391: 24 &  CG0563              &  139  & 144.6  & 74.4   &   320.5  & 1.19  & 2.35  \\
392: 25 &  CG0752              &  90   & 47.3   & 45.1   &  105.3   & 1.46  & 2.35  \\
393: 26 &  SBS1533+574         &  47   & 11.3   & 4.7    &    9.6   & 1.10  & 1.93  \\
394: 27 &  Mrk1499             &  39   & 2.7    & 2.0    &   7.6    & 1.35  & 2.46  \\
395: 28 &  Mrk930              &  77.5 & 88.4   & 40.5   &  $<$129.0& 1.14  &$<$2.17\\
396: 
397: \enddata
398: 
399: \tablenotetext{a}{The distances to the galaxies of the sample are
400:   adopted from \citet{Moustakas06} when available, while the rest of
401:   the sources are calculated from the redshifts taken from NED,
402:   assuming a $\Lambda$CDM cosmology with
403:   H$_0$\,=\,70~km~s$^{-1}$\,Mpc$^{-1}$, $\Omega_m$\,=\,0.3 and
404:   $\Omega_{\lambda}\,=\,0.7$. For IZw18, we adopt the newly derived
405:   distance by \citet{Aloisi07}. The average uncertainty in the
406:     distances is $\sim$5\%, mainly due to the value of H$_0$.}
407: \tablenotetext{b}{This is calculated based on the ``equivalent'' {\em
408:     IRAS} 60 and 100\,$\mu$m fluxes (see Section 3.4).}
409: 
410: \end{deluxetable*}
411: 
412: Using the data from this sample of low metallicity dwarf galaxies we
413: plot in Fig. 1 the radio luminosity of the sample as a function of the
414: FIR luminosity.  The luminosities of the sources we study span nearly
415: 4 orders of magnitudes, but the correlation between the FIR and the
416: radio is remarkably tight.  The scatter in the ratio of the FIR and
417: radio luminosities is 0.23 dex, i.e. less than a factor of 2. We
418: performed a least-squares bisector fit to the data and found:
419: log[L$_{\rm 1.4GHz}$(WHz$^{-1}$)]=1.09$\times$log(L$_{\rm
420:   FIR}$/L$_\odot$)+10.75 .  For comparison, we have included in the
421: plot the starburst galaxies from \citet{Brandl06} marked with
422: triangles, which has an identical slope (within 1$\sigma$).  This
423: slope of 1.09$\pm$0.07 for the dwarf galaxy data agrees well with the
424: slope of 1.10$\pm$0.04 found by \citet{Bell03} for a sample of 162
425: galaxies, as well as the slope of 1.11$\pm$0.02 for the infrared
426: selected sources from the {\em IRAS} Bright Galaxy Sample (BGS)
427: \citep{Condon91}.  This is also in agreement with \citet{Hopkins02}
428: and would suggest that globally our BCDs have a very similar FIR/radio
429: correlation to normal galaxies.
430: 
431: 
432: 
433: 
434: Another way to parameterize the IR/radio correlation is to calculate
435: the ratio of FIR to radio luminosity $q_{\rm FIR}$ according to the
436: \citet{Helou85} formula as well as the $q_{24}$ following the
437: definition of \citet{Appleton04}\footnote{We use q$_{\rm
438:     FIR}$=log[1.26$\times 10^{-14}$ (2.58S$_{60\mu m}$+S$_{100\mu
439:     m}$)/(3.75 $\times 10^{12}$ F$_{1.4 \rm GHz})$] where S$_{60\mu
440:     m}$ and S$_{100\mu m}$ are in Jy and F$_{1.4 \rm GHz}$ is in
441:   Wm$^{-2}$Hz$^{-1}$ \citep{Helou85}. We also define $q_{24}$= log
442:   (S$_{24 \mu m}$/S$_{1.4 \rm GHz}$) where S$_{24 \mu m}$ and S$_{1.4
443:     \rm GHz}$ are in Jy as in \citet{Appleton04}.}. We plot the
444: $q_{24}$ values for our sample as a function of the 24$\mu$m
445: luminosity of the galaxies in Fig. 2. For this sample we find that
446: $q_{\rm FIR}=$ 2.4$\pm$0.2, consistent with the value of normal
447: galaxies of $q_{\rm FIR}=$ 2.3$\pm$0.2 found by \citet{Condon92}. When
448: using the mid-IR 24$\mu$m data, we find that $q_{24}=$1.3$\pm$0.4 (see
449: Table \ref{tab2}). The standard deviation in $q_{24}$ is $\sim$twice
450: that of $q_{\rm FIR}$. This is not unexpected given that the spectrum
451: of star forming galaxies shows substantially larger variations in
452: spectral slope in the mid-IR \citep[see][]{Brandl06} than in the FIR
453: \citep{Dale06}. A small change in the geometry of the emitting regions
454: would affect F$_{\nu}$(24\,$\mu$m) and thus q$_{24}$ ratio much more
455: than the FIR emission and q$_{\rm FIR}$.  A similar result has also
456: been noticed by \citet{Appleton04} and \citet{Murphy06a} who found a
457: larger dispersion in q$_{24}$ as compared to q$_{70}$\footnote{We
458:   define q$_{70}$= log (S$_{70 \mu m}$/S$_{1.4 \rm GHz}$) where S$_{70
459:     \mu m}$ and S$_{1.4 \rm GHz}$ are in Jy as in \citet{Appleton04}.}
460: and suggested that this is probably due to a larger intrinsic
461: dispersion in the IR/radio correlation at shorter wavelengths.
462: Interestingly, the \citet{Appleton04} values for q$_{24}=$
463: 0.84$\pm$0.28 or the k-corrected q$_{24}$ of 0.94$\pm$0.23 are
464: somewhat smaller than our results, though consistent within 2$\sigma$.
465: One possible explanation is that the dust temperature of low
466: luminosity dwarf galaxies tends to peak at shorter wavelength than
467: normal star forming galaxies, which would result in an elevated
468: 24\,$\mu$m based luminosity. An alternative explanation is that due to
469: the peculiar morphology of dwarf galaxies, more electrons escape from
470: the galaxy because of the cosmic ray diffusion.  \citet{Boyle07} have
471: also found a high q$_{24}$=1.39$\pm$0.02 for the sources they studied,
472: and postulated that this may be due to a change in the mean q$_{24}$
473: ratio for objects with F$_\nu$(24\,$\mu$m) $<$1\,mJy, however, their
474: sources are much further away and may not be comparable to the low
475: metallicity star forming galaxies we study in the local universe.
476: 
477: \begin{figure}
478:   \epsscale{1.2}
479:   \plotone{f2.eps}
480:   \caption{The $q_{24}$ as a function of the the 24\,$\mu$m
481:     monochromatic luminosity L$_{24}$. The symbols are
482:     the same as in Fig. 1.}
483:   \label{fig:fig2}
484: \end{figure}
485: 
486: 
487: \subsection{Metallicity and Dust Temperature Effects on q$_{IR}$}
488: 
489: A number of physical parameters, such as the metallicity, dust grain
490: size distribution, and temperature may affect the shape of the
491: infrared SED. Consequently, we search for correlations between the q
492: ratios of the galaxies in our sample with those parameters.  Our
493: sample covers a metallicity range of 7.2$\le$12+log(O/H)$\le$8.9. In
494: low metallicity galaxies, the dust content is usually low, even though
495: there are notable exceptions such as SBS0335-052E \citep{Houck04b},
496: thus the emission of UV light might not be fully reprocessed by the
497: dust and re-emitted at the infrared wavelengths.  However, these
498: galaxies might have a quenched synchrotron radiation. This can be
499: attributed to various reasons including a lack of supernovae remnants
500: which accelate particles producing radio emission, the escape of fast
501: cosmic rays from the galaxy, etc.  These two competing factors, dust
502: and radio emission, counter balance each other \citep[see][]{Bell03}.
503: BCDs though, are typically small, less than $\sim$1\,kpc in size, and
504: have fewer HII regions with massive stars (w.r.t. normal galaxies)
505: which heat the dust and go supernovae to produce cosmic rays. As a
506: result the averaging effects in the sampling of the properties of the
507: interstellar medium which result from the limited spatial resolution
508: in dense galactic disks is not so prominent in BCDs. Furthermore, the
509: small number of statistics may also contribute to the observed higher
510: dispersion in q$_{24}$ ratios in metal-poor dwarf galaxies.  We
511: investigate how metallicity affects the q$_{24}$ ratios by dividing
512: the dwarf galaxies in our sample into two groups: a lower metallicity
513: group with 12+log(O/H)$\le$8.0 and a higher metallicity group with
514: 12+log(O/H)$>$8.0. This metallicity threshold was selected following
515: \citet{Rosenberg07} who noticed a change in the properties of the
516: star-forming dwarf galaxies they studied around this metallicity
517: value. We find a mean q$_{24}= $1.1$\pm$0.1\footnote{Here we use the
518:   standard error of the mean (SEM) to quantify the dispersion in the
519:   mean value.} for the first group and q$_{24}=$ 1.3$\pm$0.1 for the
520: second one.
521: %, which agree with each other within 2$\sigma$. 
522: If we calculate the q$_{24}$ for the low
523: metallicity group without including SBS0335-052E, we find a much lower
524: q$_{24}=$ 0.9$\pm$0.1. 
525: %This makes the difference in the mean q$_{24}$ 
526: %ratios between the low and high metallicity galaxies as much as 
527: %4$\sigma$.
528: %This is still consistent with the mean q$_{24}$ within 2$\sigma$.
529: %In subsection 3.4, we discuss the two
530: %lowest metallicity sources in our sample, IZW18 and SBS033-052E, which
531: %display very different q$_{24}$ ratios.  
532: We also notice that it appears q$_{24}$ generally decreases with
533: reduced metallicity for sources with 12+log(O/H)$<$8.0, with
534: SBS035-052E as a clear outlier, and the slope flattens out at
535: 12+log(O/H)$>$8.0 (see Fig. 3). This scatter in q$_{24}$ at higher
536: metallicities could be attributed to the more dispersion in the mid-IR
537: emission in these sources and is consistent with a number of other
538: studies that have found increased dispersion in the q ratios for very
539: high luminosity galaxies (which usually have higher metallicities)
540: \citep{Condon91, Yun01,Bell03}.
541: 
542: \begin{figure}
543:   \epsscale{1.2}
544:   \plotone{f3.eps}
545:   \caption{The q$_{24}$ ratio plotted as a function of the oxygen
546:     abundance of the BCDs. The mean q$_{24}$ value for all the sources
547:     is indicated by the solid line. A fit to the low metallicity
548:     sources (12+log(O/H)$\le$8.0) is indicated by the dashed line;
549:     while the dotted line is the same fit excluding SBS0335-052E. }
550:   \label{fig:fig3}
551: \end{figure}
552: 
553: 
554: 
555: \citet{Hummel88} have found evidence that, for a given galaxy, there
556: is a small decrease in F(100\,$\mu$m)/S(20\,cm) when the dust
557: temperature increases.  The latter can be traced by the ratio of
558: F(60\,$\mu$m)/F(100\,$\mu$m). \citet{Roussel03} reached a similar
559: result for the starburst galaxies they studied, in which they detected
560: an anti-correlation of q$_{\rm FIR}$ with
561: F(60\,$\mu$m)/F(100\,$\mu$m).  However, performing the same test for
562: the q$_{24}$ ratios on our dwarf galaxy sample which spans a
563: range of flux ratios -0.4 $<$log({\rm
564:   F(60\,$\mu$m)/F(100\,$\mu$m)})$<$0.2 we find no such clear
565: trend. It is conceivable that this is partly because our sample is too
566: small to identify a trend due to the intrinsic scatter in q$_{24}$.
567: 
568: 
569: \subsection{Star Formation Rate Estimates}
570: 
571: Both the radio and infrared emission can be used to estimate the star
572: formation rates (SFRs) in galaxies.  However, these correlations depend
573: on a number of parameters including dust content, optical depth and
574: metallicity. This topic and potential caveats have been discussed
575: extensively in the literature \citep[see][ and references
576: therein]{Condon92,Kennicutt98}.  For dwarf galaxies, the topic has
577: been addressed in detail by \citet{Hopkins02}. The recent wealth of
578: data from {\em Spitzer} has also provided sufficient motivation to
579: establish a calibration for the SFR using the infrared.  \citet{Wu05}
580: and more recently \citet{Calzetti07} have explored this topic using a
581: large sample of star forming regions and nearby galaxies.
582: 
583: 
584: 
585: In Fig. 4 we plot the IR estimated SFR for our sample using the
586: calibrations proposed by \citet{Wu05} and
587: \citet{Calzetti07} respectively:
588: \begin{equation}
589: {\rm SFR_{24}}(M_{\odot}yr^{-1})=\frac{\nu L_\nu(24 \mu m)}{6.66\times10^8L_\odot}
590: \end{equation}
591: 
592: 
593: \begin{equation}
594: {\rm SFR_{24}}(M_{\odot}yr^{-1})=1.27\times 10^{-38}[L_{24}(erg s^{-1})]^{0.885}
595: \end{equation}
596: 
597: as a function of the well-known radio to SFR formula of
598: \citet{Condon92}:
599: 
600: \begin{equation}
601: {\rm SFR_{1.4GHz}}=5.5\times\frac{L_{\rm 1.4GHz}}{4.6\times10^{21}(\rm{W Hz^{-1}})}
602: \end{equation}
603: 
604: 
605: The \citet{Wu05} work is based on a global correlation without
606: separating the low metallicity sources from metal rich galaxies. As 
607: can be seen in Fig. 4, a good agreement exists between the radio and
608: IR estimated SFRs (indicated by diamonds). If we use the more recent
609: calibration by \citet{Calzetti07} on SFRs from 24\,$\mu$m
610: luminosities, we find that most of the mid-IR estimated SFRs (marked
611: as filled circles) are located below the 1:1 proportionality line, and
612: thus they are consistently lower than both the SFRs estimated from the
613: radio or from \citet{Wu05}.  This is not unexpected since eq. 2 is
614: calibrated based on the high metallicity sources that have significant
615: 24\,$\mu$m emission, while most of our sources are metal-poor galaxies.
616: We should also note that \citet{Calzetti07} measure the SFR in
617: individual H{\sc ii} regions in apertures within disks and subtract a
618: ``disk background'', which could partly explain the deviations we see
619: from our analysis.  These authors have also mentioned that the SFRs of
620: low metallicity galaxies would be underestimated by a factor of 2--4
621: depending on how metal-poor the galaxies are. This deviation from a
622: linear correlation is likely due to the lower opacities for decreasing
623: metal content \citep{Walter07}.  When we fit the filled circles on
624: Fig. 4, we find that log(SFR[IR])=0.94$\times$log(SFR[1.4GHz]/3.98), which is
625: consistent with the metallicity correction factor suggested by
626: \citet{Calzetti07}. We should stress once more though that as
627: \citet{Bell03} has noted, the good agreement between the IR and radio
628: estimates of SFRs does not necessarily mean that these are the
629: ``true'' SFRs for the dwarf galaxies we study, but rather the
630: competition of the lower dust content and suppressed synchrotron
631: emission balances each other.
632: 
633: \begin{figure}
634:   \epsscale{1.2}
635:   \plotone{f4.eps}
636:   \caption{A plot of the SFR based on the 24\,$\mu$m luminosity as a
637:     function of the corresponding SFR estimated from the 1.4\,GHz
638:     radio continuum emission for each galaxy in our sample.  The open
639:     diamonds indicate the SFRs calculated using the 24\,$\mu$m fluxes
640:     following \citet{Wu05} (eq.1), the filled circles are those using
641:     eq.2 from \citet{Calzetti07}. The solid line is the 1:1
642:     proportionality line.  The dotted line is a fit to the filled
643:     circles.}
644:   \label{fig:fig4}
645: \end{figure}
646: 
647: 
648: \subsection{The Two Extremes: IZw18 and SBS0335-052E}
649: 
650: Despite this overall agreement between the IR/radio correlation in our
651: dwarf sample and the corresponding values of normal galaxies, the two
652: lowest metallicity galaxies in our sample, IZw18 and SBS0335-052E,
653: deviate markedly from the correlations and in opposite directions (see
654: Fig. 3\footnote{In Fig. 2 the ``outlier'' status of SBS0335-052E does
655:   not appear as prominent as in Fig. 3, because its high 24\,$\mu$m
656:   luminosity places it in a region of the plot where other sources
657:   with higher metallicity and higher luminosity are located.} ). As the
658: two most well-studied BCDs, it is interesting to inspect the IR/radio
659: properties of these two galaxies and a comparison of their physical
660: parameters can be found in Table \ref{tab3}.
661: 
662: 
663: \begin{deluxetable}{llcc}
664: \tabletypesize{\scriptsize}
665: \setlength{\tabcolsep}{0.2in}
666: \tablecaption{Comparison of SBS\,0335-052E and I\,Zw\,18\label{tab3}}
667: \tablewidth{0pc}
668: \tablehead{ 
669:   \colhead{}  & \colhead{SBS\,0335-052E} & \colhead{I\,Zw\,18} & \colhead{Ratio\tablenotemark{a}}\\
670: }
671: \startdata
672: 12+log(O/H)            & 7.32    &   7.17   &  1.4  \\
673: D (Mpc)                & 58      &   18     &  3    \\
674: R (pc)                 & 560\tablenotemark{b} & 870\tablenotemark{c} & $\sim$0.6 \\
675: SFR(M$_\odot$\,yr$^{-1}$)& 7$\times$10$^{-1}$     &   5$\times$10$^{-2}$   &  14    \\
676: SNR(yr$^{-1}$)          & 6$\times$10$^{-3}$   &   2$\times$10${-4}$ &  30    \\ 
677: L$_{\rm H\alpha}$(L$_\odot$) & 5.6$\times$10$^7$\tablenotemark{d}  &  1.6$\times$10$^6$    &   35   \\
678: L$_{24\mu m}$(L$_\odot$)  & 9$\times$10$^8$   &   7$\times$10$^6$   & 120   \\
679: L$_{\rm IR}$(L$_\odot$)   & $\sim$4$\times$10$^9$   &  $\sim$4$\times$10$^7$  &  100   \\
680: L$_{\rm 1.4GHz}$(W\,Hz$^{-1}$)    & 2$\times$10$^{20}$   &  8$\times$10$^{19}$   & 2.5   \\
681: M(Ks)(mag)             & -18.3   &  -16.1   &  $\sim$80\tablenotemark{e}  \\
682: 
683: \enddata
684: \tablenotetext{a}{Ratio of the relative quantities in SBS0335-052E to IZw18.}  
685: \tablenotetext{b}{This is the size of the six super star clusters in SBS0335-052E, 
686:   which has a size of 2$\arcsec$ \citep{Thuan97}.  At a distance of 58\,Mpc, 
687:   1$\arcsec$ $\sim$280\,pc.}  
688: \tablenotetext{c}{This indicate the largest projected linear extend of the total 
689:   star-forming regions in this galaxy \citep[$\sim$10$\arcsec$,][]{Hunt05b}. At 
690:   a distance of 18\,Mpc, 1$\arcsec$ $\sim$ 87\,pc.}.
691: \tablenotetext{d}{Flux in 1$\arcsec$ slit multiplied by a factor of 2 to correct for aperture effects.}
692: \tablenotetext{e}{This is the approximate mass ratio of the two sources assuming 
693:   that their mass scales with K-band luminosity and that both sources have the 
694:   same M/L$_{\rm Ks}$ \citep[see][]{Bell01}.}
695: 
696: 
697: \end{deluxetable}
698: 
699: 
700: 
701: As indicated by the {\em Spitzer} IRS spectrum of SBS0335-052E
702: \citep{Houck04b}, the SED of the galaxy has an unusually flat mid-IR
703: slope that peaks at $\sim$28$\mu$m.  \citet{Hunt05a} used {\em DUSTY}
704: models \citep{Ivezic99} to fit its SED and found a q$_{\rm FIR}$ value
705: of 2.0 for this galaxy, lower than the typical q$_{\rm FIR}$ of
706: $\sim$2.3 expected for late type systems. These calculations are based
707: on a 1.4GHz continuum flux of $\sim$0.46\,mJy, which corresponds to
708: the compact part (6\arcsec) of SBS0335-052E \citep{Dale01, Hunt04}.
709: Similarly, if we simply calculate the q ratio independent of any
710: modelling work and only use the observable parameters, we find a
711: q$_{70}=$ 2.1$\pm$0.1 \citep[assuming that the MIPS 70\,$\mu$m flux
712: density is 51.1$\pm$4.8\,mJy,][]{Engelbracht07}. This is in good
713: agreement with the q$_{70}=$ 2.16$\pm$0.17 for the sources in the
714: First Look Survey studied by \citet{Appleton04}.
715: 
716: However, when we examine the q$_{24}$ for SBS0335-052E, adopting a
717: MIPS 24\,$\mu$m flux of 66\,mJy, we find a q$_{24}$ ratio of
718: 2.2$\pm$0.1.  We see a 2$\sigma$ excess in the q parameter compared
719: with the mean q$_{24}$ of 1.3$\pm$0.4 in the dwarf sample. If this
720: excess is real, it would be consistent with the unique shape of the
721: SED of the galaxy, which suggests that the dust grain distribution is
722: dominated by small grains with temperatures of $\sim$150K and a total
723: dust mass of $\sim10^3$ M$_{\odot}$\citep{Houck04b}.  Whether the
724: large grains were never created or they were destroyed by shocks is
725: unknown.
726: %It is the
727: %surprisingly strong 24$\mu$m emission from the SBS0335-052E that
728: %compensates for the shape of the SED and results in a normal q$_{70}$.
729: The elevated q$_{24}$ ratio could also be due to the lack of
730: synchrotron emission. \citet{Roussel03,Roussel06} have shown their
731: study of three starburst galaxies, all of which have q$_{\rm FIR}$
732: more than 3$\sigma$ higher the mean q$_{\rm FIR}$ and categorized them
733: to be nascent starbursts. SBS0335-052E though, could also be a
734: candidate of this group. It is based partly on the high q$_{24}$
735: value, as well as its low H$\alpha$ luminosity as compared to the
736: infrared or radio luminosities (see Table \ref{tab3}).  Finally,
737: \citet{Hunt04} have shown the existence of free-free absorption in the
738: radio spectrum of SBS0335-052E, which is usually caused by young, 
739: dense and heavily embedded clusters\footnote{Note, that if we could
740:   properly account for the self-absorption of the radio continuum, the
741:   intrinsic q ratio would decrease. However, the exact fraction for
742:   absorption is not known and more data at $\nu<$1.5GHz would be
743:   needed to determine that number.}.
744: 
745: IZw18, despite its similar metallicity to SBS0335-052E
746: (12+log(O/H)=7.17 and 7.32 respectively), shows a rather different q
747: ratio. It also has an IR luminosity of L$_{\rm
748:   IR}$$\sim$10$^7$L$_\odot$, just 1\% of the L$_{\rm IR}$ in
749: SBS0335-052E.  In low luminosity galaxies, radio emission is known to
750: decrease faster than dust emission \citep{Devereux89}, and high
751: IR/radio ratios have been observed in low luminosity dwarf galaxies
752: \citep{Klein84}.
753: %, which can be attributed to
754: %synchrotron deficiency, optically thick free-free emission, or rapid
755: %CR escape.  
756: In IZw18 though, we find a rather low infrared to radio ratio. The
757: q$_{24}$ is found to be 0.5$\pm$0.1, nearly 2$\sigma$ lower than the
758: average value of our sample. At longer wavelengths, IZw18 is faint and
759: no {\em IRAS} FIR data are available.  However, as noted by
760: \citet{Wu07a} (see their Fig. 5), it has a very similar 5--38$\mu$m
761: continuum slope to the typical starburst galaxy NGC7714
762: \citep{Brandl04}.  If we were to assume that this similarity extends
763: to the FIR and scale down the {\em IRAS} 60 and 100\,$\mu$m flux densities
764: of NGC7714 by a factor of 375 so that its corresponding 22\,$\mu$m
765: flux density matches the one of IZw18, we find its ``equivalent''{\em
766:   IRAS} 60 and 100\,$\mu$m flux density to be 29.8 and 32.8\,mJy
767: respectively.  This would result in q$_{\rm FIR}$ of 1.3 for the
768: galaxy, which deviates from the average q$_{\rm FIR}$ ratio of
769: 2.3$\pm$0.2 calibrated by \citet{Condon92} by nearly 5$\sigma$.  If we
770: were to use the MIPS 70\,$\mu$m detection of 34\,mJy for IZw18
771: \citep{Engelbracht07} and calculate its q ratio, we would find
772: q$_{70}$=1.4, again 5$\sigma$ away from the average q$_{70}$ suggested
773: by \citet{Appleton04}.
774: 
775: A number of plausible scenarios were considered to explain this
776: result, though none appears convincing. As discussed in \citet{Wu07a},
777: because the optical depth of IZw18 is small, it could be that a
778: significant fraction of the UV light has leaked out without being
779: absorbed by the dust, thus resulting in a damped 24\,$\mu$m emission.
780: Alternatively, IZw18 may simply have an unusually high radio
781: luminosity at 1.4\,GHz.  Could it be that IZw18 is observed at a
782: special moment right after the explosion of a supernovae event? Radio
783: supernovae fade by more than a factor of 10 within $\sim$3\,yr
784: \citep{Chevalier82}. The radio continuum observations of IZw18 span
785: over a period of more than 5\,yrs \citep{Hunt05b,Cannon05} but show no
786: variation, thus this is not likely.
787: 
788: The recent results by \citet{Murphy06b} provide another possible
789: scenario. These authors analyzed a sample of nearby spiral galaxies
790: and found that systems with higher disk-averaged SFRs ($\Sigma\,_{\rm
791:   SFR}$) have usually experienced a recent episode of enhanced star
792: formation. As a result they contain a higher fraction of young cosmic ray
793: electrons that have traveled only a few hundred parsecs from their
794: acceleration sites in supernova remnants.  This is perhaps the case
795: for IZw18. However, other sources in our sample also have elevated
796: $\Sigma\,_{\rm SFR}$ and show no radio excess. It could also be that
797: the extremely low metal abundance of IZw18, 0.15\,dex lower than that
798: of SBS0335-052E, is below a critical threshold. The newly formed
799: stars in the unpolluted interstellar medium may produce, and
800: subsequently heat, less dust than electrons which are accelerated and
801: contribute in the radio emission. We also note that the rate of SNR in
802: IZw18 is a factor of $\sim$30 lower than that in SBS0335-052E (see
803: Table \ref{tab3}) but its morphology is much more disturbed with
804: filamentary structure and outflows \citep[see][]{Izotov04}. If we were
805: to assume that both sources have similar mass-to-light ratios and
806: estimate their mass from their K-band luminosities, we find that the
807: mass of IZw18 is almost a factor of 80 less than that of SBS0335-052E.
808: Thus the rate of SNR normalized with mass in IZw18 would be $\sim$2.5
809: times that of SBS0335-052E. Could it be that the disturbed morphology
810: of IZw18 in addition to its low metal content that it created the
811: conditions in the interstellar medium for this abnormally high radio
812: flux? The question remains open.
813: 
814: 
815: \section{Conclusions}
816: 
817: We have studied the mid-IR and FIR to radio correlation in a sample of
818: dwarf star-forming galaxies spanning a metallicity range from
819: 7.2$<$12+log(O/H)$<$8.9, using {\em Spitzer} IRS/MIPS, as well as radio
820: 1.4\,GHz data obtained from the literature.  The BCD sample appears to
821: follow the same FIR/radio correlation as normal star forming galaxies.
822: The analysis based on mid-IR and radio data reveals a similar
823: correlation, though the scatter is larger, probably due to an
824: intrinsically higher variation in the 15-30$\mu$m SEDs of dwarf
825: galaxies.  When comparing the q$_{24}$ ratios with metallicity or
826: effective dust temperature, we find no strong correlation, though there
827: is a general trend of lower q ratios at lower metallicity for
828: galaxies with 12+log(O/H)$<$8.0 and the correlation flattens out
829: toward higher metallicity. In general the SFRs estimated from the
830: radio 1.4\,GHz continuum and the mid-IR data are in good agreement.
831: Two extremely metal poor BCDs, IZw18 and SBS0335-052E appear to
832: deviate from the average q$_{24}$ by $\sim$2$\sigma$, with one
833: displaying a radio excess and the other an infrared excess.
834: 
835: 
836: 
837: 
838: \acknowledgments 
839: 
840: The authors would like to thank P. Appleton and D. Calzetti for
841: stimulating discussions. We would also like to thank G. Helou, L.K.
842: Hunt as well as an anonymous referee whose detailed comments and
843: insightful suggestions have helped to improve this manuscript. This
844: work is based in part on observations made with the {\em Spitzer} Space
845: Telescope, which is operated by the Jet Propulsion Laboratory,
846: California Institute of Technology, under NASA contract 1407. Support
847: for this work was provided by NASA through Contract Number 1257184
848: issued by JPL/Caltech. VC would like to acknowledge the partial
849: support from the EU ToK grant 39965.
850: 
851: 
852: \begin{thebibliography}{}
853: \bibitem[Allende Prieto et al.(2001)]{Prieto01} 
854:   Allende Prieto, C., Lambert, D.~L., \& Asplund, M.\ 2001, \apjl, 556, L63 
855: 
856: \bibitem[Aloisi et al.(2007)]{Aloisi07}
857:   Aloisi, A. et al.\ 2007, \apjl, 667, L151
858: 
859: \bibitem[Appleton et al.(2004)]{Appleton04} 
860:   Appleton, P.~N., et al.\ 2004, \apjs, 154, 147 
861: 
862: \bibitem[Becker et al.(1995)]{Becker95} 
863:   Becker, R.~H., White, R.~L., \& Helfand, D.~J.\ 1995, \apj, 450, 559 
864: 
865: \bibitem[Bell \& de Jong(2001)]{Bell01} 
866:   Bell, E.~F., \& de Jong, R.~S.\ 2001, \apj, 550, 212
867: 
868: \bibitem[Bell(2003)]{Bell03} 
869:   Bell, E.~F.\ 2003, \apj, 586, 794 
870: 
871: \bibitem[Bergvall \& \"{O}stlin(2002)]{Bergvall02} 
872:   Bergvall, N., \& \"{O}stlin, G.\ 2002, \aap, 390, 891 
873: 
874: \bibitem[Brandl et al.(2004)]{Brandl04} 
875:   Brandl, B.~R., et al.\ 2004, \apjs, 154, 188 
876: 
877: \bibitem[Brandl et al.(2006)]{Brandl06} 
878:   Brandl, B.~R., et al.\ 2006, \apj, 653, 1129 
879: 
880: \bibitem[Boyle et al.(2007)]{Boyle07} 
881:   Boyle, B.~J., Cornwell, T.~J., Middelberg, E., Norris, R.~P., Appleton, P.~N., \& Smail, I.\ 2007, \mnras, 376, 1182 
882: 
883: \bibitem[Calzetti et al.(2007)]{Calzetti07} 
884:   Calzetti, D., et al. \apj, 666,870
885: 
886: \bibitem[Cannon et al.(2005)]{Cannon05} 
887:   Cannon, J.~M.,  Walter, F., Skillman, E.~D., \& van Zee, L.\ 2005, \apjl, 621, L21 
888: 
889: \bibitem[Chevalier(1982)]{Chevalier82} 
890:   Chevalier, R.~A.\ 1982, \apj, 258, 790 
891: 
892: \bibitem[Condon et al.(1991)]{Condon91} 
893:   Condon, J.~J., Anderson, M.~L., \& Helou, G.\ 1991, \apj, 376, 95 
894: 
895: \bibitem[Condon(1992)]{Condon92} 
896: Condon, J.~J.\ 1992, \araa, 30, 575 
897: 
898: \bibitem[Condon et al.(1998)]{Condon98} Condon, J.~J., Cotton, W.~D.,
899:   Greisen, E.~W., Yin, Q.~F., Perley, R.~A., Taylor, G.~B., \&
900:   Broderick, J.~J.\ 1998, \aj, 115, 1693
901: 
902: \bibitem[Dale et al.(2001)]{Dale01} 
903:   Dale, D.~A., Helou, G., Neugebauer, G., Soifer, B.~T., Frayer, D.~T., \& Condon, J.~J.\ 2001, \aj, 122, 1736 
904: 
905: \bibitem[Dale et al.(2006)]{Dale06} 
906:   Dale, D.~A., et al.\ 2006, \apj, 646, 161 
907: 
908: \bibitem[de Jong et al.(1985)]{Dejong85} 
909:   de Jong, T., Klein, U., Wielebinski, R., \& Wunderlich, E.\ 1985, \aap, 147, L6 
910: 
911: \bibitem[Devereux \& Eales(1989)]{Devereux89} 
912:   Devereux, N.~A., \& Eales, S.~A.\ 1989, \apj, 340, 708 
913: 
914: 
915: \bibitem[Elbaz et al.(2002)]{Elbaz02} 
916:   Elbaz, D., Cesarsky, C.~J., Chanial, P., Aussel, H., Franceschini, A., Fadda, D., \& Chary, R.~R.\ 2002, \aap, 384, 848 
917: 
918: \bibitem[Engelbracht et al.(2005)]{Engelbracht05} 
919:   Engelbracht, C.~W., Gordon, K.~D., Rieke, G.~H., Werner, M.~W., Dale, D.~A., \& Latter, W.~B.\ 2005, \apjl, 628, L29 
920: 
921: \bibitem[Engelbracht et al.(2007)]{Engelbracht07}
922:   Engelbracht, C.~W., et al.\ 2007, \apj, submitted
923: 
924: \bibitem[Garrett(2002)]{Garrett02} 
925:   Garrett, M.~A.\ 2002, \aap, 384, L19
926: 
927: \bibitem[Gruppioni et al.(2003)]{Gruppioni03} 
928:   Gruppioni, C., Pozzi, F., Zamorani, G., Ciliegi, P., Lari, C., Calabrese, E., La Franca, F., \& Matute, I.\ 2003, \mnras, 341, L1
929: 
930: \bibitem[Guseva et al.(2000)]{Guseva00} 
931:   Guseva, N.~G., Izotov,    Y.~I., \& Thuan, T.~X.\ 2000, \apj, 531, 776 
932: 
933: \bibitem[Hao et al.(2007)]{Hao07} 
934:   Hao, Lei, et al.,\ 2007, in  preparation
935: 
936: \bibitem[Harwit \& Pacini(1975)]{Harwit75} 
937:   Harwit, M., \& Pacini, F.\ 1975, \apjl, 200, L127 
938: 
939: \bibitem[Heckman et al.(1998)]{Heckman98} 
940:   Heckman, T.~M., Robert, C., Leitherer, C., Garnett, D.~R., \& van der Rydt, F.\ 1998, \apj, 503, 646 
941: 
942: \bibitem[Helou et al.(1985)]{Helou85} 
943:   Helou, G., Soifer, B.~T., \& Rowan-Robinson, M.\ 1985, \apjl, 298, L7
944: 
945: \bibitem[Helou \& Bicay(1993)]{Helou93} 
946:   Helou, G., \& Bicay, M.~D.\ 1993, \apj, 415, 93
947: 
948: \bibitem[Hopkins et al.(2002)]{Hopkins02} 
949:   Hopkins, A.~M., Schulte-Ladbeck, R.~E., \& Drozdovsky, I.~O.\ 2002, \aj, 124, 862 
950: 
951: \bibitem[Houck et al.(2004a)]{Houck04a} 
952:   Houck, J.~R., et al.\ 2004a, \apjs, 154, 18 
953: 
954: \bibitem[Houck et al.(2004b)]{Houck04b} 
955:   Houck, J.~R., et al.\ 2004b, \apjs, 154, 211 
956: 
957: \bibitem[Hummel et al.(1988)]{Hummel88} 
958:   Hummel, E., Davies, R.~D., Pedlar, A., Wolstencroft, R.~D., \& van der Hulst, J.~M.\ 1988, \aap, 199, 91 
959: 
960: \bibitem[Hunt et al.(2004)]{Hunt04} 
961:   Hunt, L.~K., Dyer, K.~K., Thuan, T.~X., \& Ulvestad, J.~S.\ 2004, \apj, 606, 853 
962: 
963: \bibitem[Hunt et al.(2005a)]{Hunt05a} 
964:   Hunt, L., Bianchi, S., \& Maiolino, R.\ 2005b, \aap, 434, 849 
965: 
966: \bibitem[Hunt et al.(2005b)]{Hunt05b} 
967:   Hunt, L.~K., Dyer, K.~K., \& Thuan, T.~X.\ 2005a, \aap, 436, 837 
968: 
969: \bibitem[Izotov \& Thuan(2004)]{Izotov04} 
970:   Izotov, Y.~I., \& Thuan, T.~X.\ 2004, \apj, 616, 768 
971: 
972: 
973: \bibitem[Izotov \& Thuan(1998)]{Izotov98} 
974:   Izotov, Y.~I., \& Thuan, T.~X.\ 1998, \apj, 500, 188 
975: 
976: \bibitem[Izotov \& Thuan(1999)]{Izotov99} 
977:   Izotov, Y.~I., \& Thuan, T.~X.\ 1999, \apj, 511, 639 
978: 
979: \bibitem[Ivezi\'c et al.(1999)]{Ivezic99} 
980:   Ivezi\'c, Z., Nenkova, M., \& Elitzur,M. 1999, User Manual for {\em DUSTY}, 
981:   University of Kentucky Internal Report,accessible at {\tt http://www.pa.uky.edu/~moshe/dusty} 
982: 
983: \bibitem[Jannuzi \& Dey(1999)]{Jannuzi99} 
984:   Jannuzi, B.~T.,~\& Dey, A.\ 1999, ASP Conf.~Ser.~191:
985:   Photometric Redshifts and the Detection of High Redshift Galaxies, 111
986: 
987: \bibitem[Kennicutt(1998)]{Kennicutt98} 
988:   Kennicutt, R.~C., Jr.\ 1998, \araa, 36, 189 
989: 
990: 
991: \bibitem[Kirshner et al.(1981)]{Kirshner81} 
992:   Kirshner, R.~P., Oemler, A., Jr., Schechter, P.~L., \& Shectman, S.~A.\ 1981, \apjl, 248, L57 
993: 
994: \bibitem[Klein et al.(1991)]{Klein91} 
995:   Klein, U., Weiland, H., \& Brinks, E.\ 1991, \aap, 246, 323 
996: 
997: \bibitem[Klein et al.(1984)]{Klein84} 
998:   Klein, U., Wielebinski, R., \& Thuan, T.~X.\ 1984, \aap, 141, 241 
999: 
1000: \bibitem[Kobulnicky \& Skillman(1997)]{Kobulnicky97} 
1001:   Kobulnicky, H.~A., \& Skillman, E.~D.\ 1997, \apj, 489, 636 
1002: 
1003: \bibitem[Kobulnicky \& Johnson(1999)]{Kobulnicky99} 
1004:   Kobulnicky, H.~A., \& Johnson, K.~E.\ 1999, \apj, 527, 154 
1005: 
1006: \bibitem[Kunth \& Joubert(1985)]{Kunth85} 
1007:   Kunth, D., \& Joubert, M.\ 1985, \aap, 142, 411 
1008: 
1009: \bibitem[Leroy et al.(2005)]{Leroy05} 
1010:   Leroy, A., Bolatto, A.~D., Simon, J.~D., \& Blitz, L.\ 2005, \apj, 625, 763 
1011: 
1012: \bibitem[Miller \& Owen(2001)]{Miller01} 
1013:   Miller, N.~A., \& Owen, F.~N.\ 2001, \aj, 121, 1903 
1014: 
1015: \bibitem[Moshir \& et al.(1990)]{Moshir90} 
1016:   Moshir, M., \& et al.\ 1990, IRAS Faint Source Catalogue, version 2.0 (1990)
1017: 
1018: \bibitem[Moustakas \& Kennicutt(2006)]{Moustakas06} 
1019:   Moustakas, J., \& Kennicutt, R.~C., Jr.\ 2006, \apjs, 164, 81 
1020: 
1021: \bibitem[Murphy et al.(2006a)]{Murphy06a} 
1022:   Murphy, E.~J., et  al.\ 2006a, \apj, 638, 157
1023: 
1024: \bibitem[Murphy et al.(2006b)]{Murphy06b} 
1025:   Murphy, E.~J., et  al.\ 2006b, \apjl, 651, L111 
1026: 
1027: \bibitem[Popescu \& Hopp(2000)]{Popescu00} 
1028:   Popescu, C.~C., \& Hopp, U.\ 2000, \aaps, 142, 247
1029: 
1030: \bibitem[Rieke et al.(2004)]{Rieke04} 
1031:   Rieke, G.~H., et al.\ 2004, \apjs, 154, 25 
1032: 
1033: \bibitem[Rosenberg et al.(2006)]{Rosenberg06} 
1034:   Rosenberg, J.~L., Ashby, M.~L.~N., Salzer, J.~J., \& Huang, J.-S.\
1035:   2006, \apj, 636, 742
1036: 
1037: \bibitem[Rosenberg et al.(2007)]{Rosenberg07} Rosenberg, J.~L., Wu,
1038:   Yanling, Le Floc'h, E., Charmandaris, V., Ashby, M.~L.~N., Houck,
1039:   J.~R., Salzer, J.~J. Willner, S.P.\ 2008, \apj, in press, astro-ph/0710.5514
1040: 
1041: \bibitem[Roussel et al.(2003)]{Roussel03} 
1042:   Roussel, H., Helou, G., Beck, R., Condon, J.~J., Bosma, A., Matthews, K., \& Jarrett, T.~H.\ 2003, \apj, 593, 733 
1043: 
1044: \bibitem[Roussel et al.(2006)]{Roussel06} 
1045:   Roussel, H., et al.\ 2006, \apj, 646, 841 
1046: 
1047: \bibitem[Sanders et al.(2003)]{Sanders03} 
1048:   Sanders, D.~B., Mazzarella, J.~M., Kim, D.-C., Surace, J.~A., \& Soifer, B.~T.\ 2003, \aj, 126, 1607 
1049: 
1050: \bibitem[Sanders et al.(2007)]{Sanders07} 
1051:   Sanders, D.~B., et al.\ 2007, \apjs, (in press arXiv:astro-ph/0701318)
1052: 
1053: \bibitem[Shi et al.(2005)]{Shi05} 
1054:   Shi, F., Kong, X., Li, C., \& Cheng, F.~Z.\ 2005, \aap, 437, 849 
1055: 
1056: \bibitem[Storchi-Bergmann et al.(1994)]{Storchi94} 
1057:   Storchi-Bergmann, T., Calzetti, D., \& Kinney, A.~L.\ 1994, \apj, 429, 572 
1058: 
1059: \bibitem[Thuan \& Izotov(2005)]{Thuan05} 
1060:   Thuan, T.~X., \& Izotov, Y.~I.\ 2005, \apjs, 161, 240 
1061: 
1062: \bibitem[Thuan et al.(1997)]{Thuan97} 
1063:   Thuan, T.~X., Izotov, Y.~I., \& Lipovetsky, V.~A.\ 1997, \apj, 477, 661 
1064: 
1065: \bibitem[van der Kruit(1971)]{van71} 
1066:   van der Kruit, P.~C.\ 1971, \aap, 15, 110 
1067: 
1068: 
1069: \bibitem[Walter et al.(2007)]{Walter07} 
1070:   Walter, F., et al.\ 2007, \apj, 661, 102 
1071: 
1072: \bibitem[Werner et al.(2004)]{Werner04} 
1073:   Werner, M.~W., et al.\ 2004, \apjs, 154, 1 
1074: 
1075: \bibitem[Wu et al.(2005)]{Wu05} 
1076:   Wu, H., Cao, C., Hao, C.-N., Liu, F.-S., Wang, J.-L., Xia, X.-Y., Deng, Z.-G., \& Young, C.~K.-S.\ 2005, \apjl, 632, L79 
1077: 
1078: \bibitem[Wu et al.(2006)]{Wu06} 
1079:   Wu, Y., Charmandaris, V., Hao, L., Brandl, B.~R.,
1080:   Bernard-Salas, J., Spoon, H.~W.~W., \& Houck, J.~R.\ 2006, \apj,
1081:   639, 157
1082: 
1083: \bibitem[Wu et al.(2007a)]{Wu07a}
1084:   Wu, Y., et al., 2007a, \apj, 662, 952
1085: 
1086: \bibitem[Wu et al.(2007b)]{Wu07b} 
1087:   Wu, Y.,  Bernard-Salas, J., Charmandaris,V. Lebouteiller, V., Hao,
1088:   L., Brandl, B.R., Houck, J.R. \ 2007b, ApJ in press, 
1089:   astro-ph/0710.0003
1090: 
1091: \bibitem[Yun et al.(2001)]{Yun01} 
1092:   Yun, M.~S., Reddy, N.~A., \& Condon, J.~J.\ 2001, \apj, 554, 803 
1093: 
1094: \end{thebibliography}
1095: 
1096: 
1097: 
1098: 
1099: 
1100: 
1101: 
1102: 
1103: 
1104: 
1105: 
1106: \end{document}
1107: 
1108: 
1109: