1: %\documentclass[preprint,12pt]{aastex}
2: %\documentclass[manuscript]{aastex}
3: \documentclass[preprint2]{aastex}
4: \usepackage{rotating,times,pictex,graphicx,latexsym,color,pifont}
5: %\usepackage{supertabular,longtable}
6: \usepackage{longtable}
7: %
8:
9: \shorttitle{Which are the youngest protostars?}
10: \shortauthors{D.~Froebrich}
11:
12: \begin{document}
13:
14: \title{Which are the youngest protostars? \\ Determining properties of
15: confirmed and candidate Class\,0 sources by broad-band photometry}
16:
17: \author{D.~Froebrich}
18: \affil{Dublin Institute for Advanced Studies, 5 Merrion Square, Dublin 2,
19: Ireland}
20:
21: \begin{abstract}
22:
23: We searched the literature to obtain a complete list of known Class\,0 sources.
24: A list of 95 confirmed or candidate objects was compiled. To the best of our
25: knowledge, all published broad-band observations from 1\,$\mu$m to 3.5\,mm have
26: been collected and are assembled in a catalogue. These data were used to
27: determine physical properties (T$_{\rm bol}$, L$_{\rm bol}$, L$_{\rm
28: smm}$/L$_{\rm bol}$, M$_{\rm env}$) and for a uniform classification. 50
29: sources possess sufficient observational data and are classified as Class\,0 or
30: Class\,0/1 objects. The source properties are compared with different
31: evolutionary models to infer ages and masses, and their correlations are
32: investigated. About 25\% of the sources are found to be in a quiet accretion
33: phase or possess a significantly different time evolution of the accretion rate
34: than the average. In Taurus, with its isolated star formation mode, this seems
35: especially to be the case.
36:
37: \end{abstract}
38:
39: \keywords{Catalogs -- Stars: evolution -- Stars: formation -- Infrared: stars}
40:
41: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
42: %%%%%%%%%%%%%%%%%%%%%% Introduction %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
43: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
44:
45: \section{Introduction}
46:
47: Protostars begin their life soon after the collapse of their parental cloud
48: core begins. Such stellar embryos have masses of less than 10$^{-2}$\,M$_\odot$
49: (Larson \cite{l03}) and are deeply embedded in a dense and massive envelope of
50: gas and dust. The central object grows in mass by accreting material from this
51: envelope either by direct infall or via a flattened circumstellar disc. As long
52: as the central object is less massive than the surrounding envelope it is
53: called a Class\,0 object. Otherwise it is a Class\,1 protostar. Since this mass
54: ratio is not directly observable, other criteria have to be adopted for a
55: classification. Andr\'e et al. \cite{2000prpl.conf...59A} defined three
56: observational properties needed to classify an object as a Class\,0 source: (i)
57: an internal heating source (compact centimeter radio continuum) or bipolar
58: outflow as an indication of a central young stellar object; (ii) extended and
59: centrally peaked sub-millimeter continuum emission, indicating the spheroidal
60: envelope; (iii) a high ratio of sub-millimeter to bolometric luminosity
61: (L$_{\rm smm}$/L$_{\rm bol}$\,$>$\,0.005), with L$_{\rm smm}$ measured longward
62: of 350\,$\mu$m. It is shown in Andr\'{e} et al. \cite{1993ApJ...406..122A} that
63: this last criterion is equivalent to the mass ratio of the envelope and the
64: central star being larger than one. Other authors use the bolometric
65: temperature (T$_{\rm bol}$) instead of L$_{\rm smm}$/L$_{\rm bol}$ (e.g. Chen
66: et al. \cite{1995ApJ...445..377C, 1997ApJ...478..295C} use T$_{\rm
67: bol}$\,$<$\,70\,K).
68:
69: The Class\,0 stage can be characterised as the main mass accretion phase, in
70: which the forming star gains the bulk of its final mass. This process is not
71: yet well understood. How does the mass accretion rate depend on the age and the
72: properties of the cloud, and is it governed by turbulence or ambipolar
73: diffusion? What is the lifetime of the Class\,0 sources? How are the source
74: properties connected to the molecular outflow, that is inevitably driven by
75: these objects? The answers to these questions will finally help us to
76: understand the feedback of the star formation process to the surrounding
77: molecular cloud and the initial mass function.
78:
79: To seriously address these questions we need a large, homogeneously classified
80: sample of these very young protostellar objects. Several samples have been
81: compiled in the recent years (e.g. Chen et al. \cite{1995ApJ...445..377C},
82: Andr\'e et al. \cite{2000prpl.conf...59A}, Shirley et al.
83: \cite{2000ApJS..131..249S}, Motte \& Andr\'{e} \cite{2001A&A...365..440M}). All
84: these works have specific problems: (1) they concentrate only on a limited
85: number of sources; (2) different criterias for classification of Class\,0
86: objects are used; (3) only parts of the available observational data are
87: considered.
88:
89: The first step to obtain a large, homogeneously classified sample of Class\,0
90: sources is a complete search of the literature for such objects. Here we
91: present an as far as possible complete list of the youngest known protostars to
92: date. We will classify them by means of their spectral energy distribution
93: (SED) using all (to the best of our knowledge) published broad-band photometric
94: data. We derive basic properties of these sources (T$_{\rm bol}$, L$_{\rm
95: bol}$). The sub-mm data are used to estimate envelope masses (M$_{\rm env}$)
96: and sub-mm slopes of the SED ($\beta$). By comparing T$_{\rm bol}$, L$_{\rm
97: bol}$, and M$_{\rm env}$ with evolutionary models we determine ages and masses
98: of these sources and investigate limitations of the current observational data
99: and evolutionary models. This is of particular interest for forthcoming
100: powerful telescopes like Spitzer and ALMA, which will enable us to improve our
101: knowledge of these sources significantly. Given this expected growth in
102: observational data in the near future, the present catalogue might be quickly
103: outdated. Hence, all the data will be available on a
104: webpage\footnote{http://www.dias.ie/protostars} and updated regularly.
105:
106: In Sect.\,\ref{sample} we introduce the source sample. The method of our data
107: analysis is then put forward in Sect.\,\ref{analysis}, including a description
108: of the source classification in Sect.\,\ref{classification}. Finally the
109: results and implications are discussed in Section\,\ref{discussion}.
110:
111:
112: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
113: %%%%%%%%%%%%% Source Sample %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
114: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
115:
116: \section{Source sample and broad band photometry}
117:
118: \label{sample}
119:
120: To obtain an as far as possible complete list of known Class\,0 objects we
121: combined the samples of candidates and sources mentioned in various
122: publications. In particular we combined: (1) sources with T$_{\rm
123: bol}$\,$<$\,100\,K from Chen et al. \cite{1995ApJ...445..377C}; (2) objects
124: with T$_{\rm dust}$\,$<$\,100\,K from Hurt et al. \cite{1996ApJ...460L..45H};
125: (3) Class\,0 sources listed in Bontemps et al. \cite{1996A&A...311..858B}; (4)
126: objects of type CL0 in Saraceno et al. \cite{1996A&A...309..827S}; (5)
127: candidates with T$_{\rm bol}$\,$<$\,100\,K from Chen et al.
128: \cite{1997ApJ...478..295C}; the source B\,35 could not be identified in SIMBAD;
129: (6) Objects with L$_{\rm bol}$/L$_{\rm smm}$\,$<$\,200 in Chini et al.
130: \cite{1997ApJ...474L.135C} with VLA detected counterparts in Reipurth et al.
131: \cite{1999AJ....118..983R}; (7) the Class\,0 sources in CB\,68 and CB\,232
132: mentioned in Huard et al. \cite{1999ApJ...526..833H}; (8) objects classified as
133: Class\,0 in Park et al. \cite{1999ApJ...520..223P}; (9) all objects listed in
134: Andr\'{e} et al. \cite{2000prpl.conf...59A}; (10) Class\,0 sources in Shirley
135: et al. \cite{2000ApJS..131..249S}; (11) sources indicated as Class\,0 or
136: Class\,0/1 in Motte \& Andr\'{e} \cite{2001A&A...365..440M}; (12) objects with
137: L$_{\rm FIR}$/L$_{\rm smm}$\,$<$\,200 in Chini et al.
138: \cite{2001A&A...369..155C}; (13) candidates with T$_{\rm bol}$\,$<$\,100\,K in
139: Lehtinen et al. \cite{2001A&A...367..311L}; (14) Class\,0 sources from Visser
140: et al. \cite{2002AJ....124.2756V}; (15) objects with T$_{\rm bol}$\,$<$\,100\,K
141: listed in Young et al. \cite{2003ApJS..145..111Y}; (16) Class\,0 objects and
142: candidates investigated in Froebrich et al. \cite{2003MNRAS.346..163F}; (17)
143: candidates and confirmed Class\,0 sources from Rengel et al.
144: \cite{2004A&A..inprep.R}; (18) the possible Class\,0 sources NGC\,7129\,FIRS2
145: (Eiroa et al. \cite{1998A&A...335..243E}), NGC\,2068\,LBS\,17 (Gibb \& Little
146: \cite{2000MNRAS.313..663G}), IC\,348\,MMS (Eisl\"offel et al.
147: \cite{2003ApJ...595..259E}), NGC\,7538\,S (Sandell et al.
148: \cite{2003ApJ...590L..45S}), and MonOB1\,IRAS12\,S1 (Wolf-Chase et al.
149: \cite{2003MNRAS.344..809W}). All investigated sources with their positions,
150: adapted distances, and references are listed in Appendix\,\ref{sourcesample} in
151: Table\,\ref{sources}. This table will be only available in electronic form and
152: lists the object name used in this paper and other common names of the sources.
153:
154: Sources in the combined sample were identified by their position in the SIMBAD
155: database in order to avoid confusion due to unclear or conflicting
156: nomenclature. A search in the related publications given by SIMBAD for the
157: sources was performed and to our best knowledge all available broad-band
158: photometric data were extracted. We restricted the search for data ranging from
159: the near infrared (NIR) to wavelengths shorter than 3.5\,mm. At longer
160: wavelengths the emission of the envelope might be a combination of dust
161: continuum and free-free emission from the central source and may influence the
162: analysis of the envelope properties. NIR and mid-infrared detections with
163: brightness given in magnitudes were converted into Jansky using flux
164: zero-points of Wamsteker \cite{1981A&A....97..329W}. In case of non-existing
165: NIR detections in the literature we searched the 2MASS catalogue. If the source
166: was not detected by 2MASS, we determined an upper limit (the faintest flux in
167: the 2MASS catalogue within 200\arcsec\, of the source).
168:
169: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
170: %%%%%%%%%%%%%%%%%%%%%% Models %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
171: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
172:
173: \section{Data analysis}
174:
175: \label{analysis}
176:
177: \subsection{Basic parameters}
178:
179: The observed SEDs have to be characterised in order to be able to compare them
180: and learn about the object properties. We computed T$_{\rm bol}$ following the
181: method described in Chen et al. \cite{1995ApJ...445..377C}. This is the
182: temperature of a black body, which posseses the same mean frequency as the
183: respective source. The bolometric luminosity is determined by integrating the
184: SED and adapting a distance. Note that L$_{\rm bol}$ might be underestimated by
185: this method, when the emission maximum is not well covered (no data from 100 to
186: 350\,$\mu$m). We compute the ratio of bolometric to sub-mm ($\lambda >
187: 350$\,$\mu$m) luminosity. This ratio is used in the source classification (see
188: Sec.\,\ref{classification}).
189:
190: The mean spectral index $\alpha$ was determined by a powerlaw fit to fluxes
191: measured longward of 400\,$\mu$m. $\alpha$ is converted to the powerlaw index
192: of the dust emissivity (assuming optically thin emission) by
193: $\beta$\,=\,$\alpha$-2. The powerlaw fit allows for a uniform determination of
194: the flux density at 1.3\,mm, even if the object was not observed at this
195: wavelength. The fluxes are scaled to a distance of 300\,pc and transformed into
196: envelope masses by M$_{\rm env}$\,=\,1.5\,M$_\odot$\,*\,F$_{\rm 1.3mm}^{\rm
197: 300pc}$[Jy] (following Motte and Andr\'e \cite{2001A&A...365..440M}). We used a
198: dust opacity at 1.3\,mm of 0.01\,cm$^2$\,g$^{-1}$ and a dust temperature of
199: 24\,K for all objects. A lower dust temperature of 15\,K, as e.g. suggested by
200: Motte \& Andr\'e \cite{2001A&A...365..440M} for the Taurus sources, would lead
201: to envelope masses which are a factor of 1.6 higher.
202:
203: The data for some sources are inhomogeneous. Different aperture sizes are used
204: at different wavelengths. To ensure a consistent treatment of the data for all
205: sources, datapoints were selected or excluded manually before the determination
206: of the source properties. In Table\,\ref{notes}
207: (Appendix\,\ref{notesonsources}, available online only) we list datapoints not
208: used for the determinations, together with further notes. The obtained
209: parameters for all our sources are listed in Table\,\ref{param}. This table is
210: divided in three sections. First we list all sources where the SED is very well
211: sampled and all parameters can be determined accuratly. The second part
212: contains sources where the SED is only well determined on one side of the peak
213: and for the other side only upper limits (e.g. VLA\,1623) or very few data
214: (e.g. CB\,232) are available, hence mostly limits for the source properties
215: could be derived. In the last part we list sources that have very insufficient
216: data and where at best only $\beta$ and M$_{\rm env}$ or a lower limit for
217: L$_{\rm bol}$ can be deterimed. The objects Trifid-TC\,3, GF\,9-2, and
218: IC\,348\,MMS are not listed in this table since none of the source properties
219: could be determined.
220:
221: \subsection{Source Classification}
222:
223: \label{classification}
224:
225: To classify the investigated objects we conducted three different tests for
226: objects listed in the first two parts of Table\,\ref{param}. A source was
227: negatively tested to be a Class\,0 object when: (1) T$_{\rm bol}$ is larger
228: than 80\,K; (2) L$_{\rm smm}$/L$_{\rm bol}$ is less than 0.005; (3) There is a
229: NIR ($\lambda$\,$<$5\,$\mu$m) detection. If all three tests are positive then
230: the source was marked as Class\,0 object (0). If at least two tests are
231: positive and the third could not be conducted or was negative, then the source
232: is marked as borderline object (0/1). In case of two negative tests the source
233: is classified as Class\,1 (1). For all remaining objects a classification
234: cannot be performed and the source is marked by '?'. In principle we further
235: need to test if the object possesses an extended envelope. Hence, we marked all
236: sources with a $^\dagger$ symbol, where non or no conclusive sub-mm or
237: millimeter map is available. The lack of such data mostly effects southern
238: hemisphere objects, which anyway could not be properly classified by our tests.
239:
240: The classification obtained for each source according to these criteria is
241: listed in the last column of Table\,\ref{param}. Positions in the T$_{\rm
242: bol}$-L$_{\rm bol}$ diagram of all sources properly classified as Class\,0 or
243: Class\,0/1 can be seen in Fig.\,\ref{t_l} together with the T$_{\rm
244: bol}$-M$_{\rm env}$ positions. In order to obtain a well defined sample with as
245: few as possible selection effects, we excluded sources with distances larger
246: than 500\,pc in the diagram.
247:
248: \begin{figure}[t]
249: \caption{\label{t_l} Distribution of L$_{\rm bol}$ (filled circles) and M$_{\rm
250: env}$ (open circles) against T$_{\rm bol}$ of the Class\,0
251: and Class\,0/1 sources with distances of less than 500\,pc.}
252: \includegraphics[width=7.cm, height=5.5cm, bb=90 350 310 520]{f1.eps}
253: \end{figure}
254:
255: \subsection{Evolutionary Models}
256:
257: \label{evolution}
258:
259: Bolometric temperature, luminosity and envelope mass are three observational
260: quantities available for these young sources. To investigate physical
261: properties, such as final masses and ages, we need to compare these values with
262: evolutionary models. For the comparison we selected all sources from part one
263: in Table\,\ref{param} and the objects from part two with a clear classification
264: (0, 0/1, or 1), and where T$_{\rm bol}$, L$_{\rm bol}$, and M$_{\rm env}$ could
265: be determined.
266:
267: There are several evolutionary schemes available in the literature. We use
268: three different models to estimate ages and final masses of our sources. (1)
269: The evolutionary diagram presented in Fig.\,12 of Myers et al.
270: \cite{1998ApJ...492..703M}, where the infall rate matches the isothermal sphere
271: infall solution at early times and exponentially declines later. (2)
272: Evolutionary tracks shown in Fig.\,6b in Andr\'e et al.
273: \cite{2000prpl.conf...59A}, where accretion rate and envelope mass decline
274: exponentially. (3) The evolutionary scheme developed by Smith
275: \cite{1998Ap&SS.261..169S, 2000IrAJ...27...25S, s02}, where the mass accretion
276: rate first increases exponentially and then shows a powerlaw fall off.
277:
278: All of these evolutionary models are subject of shortcommings. They are based
279: on assumptions about the time dependence of the mass accretion rates. While the
280: overall decline with time has substantial observational support (Calvet et al.
281: \cite{2000prpl.conf..377C}), the particular behaviour is unclear. Analytical
282: and numerical models reproduce both, exponential decline (Shu et al.
283: \cite{2004ApJ...601..930S}) or well defined early peaks followed by a fall off
284: (Schmeja \& Klessen \cite{2004A&A...419..405S}). The usage of T$_{\rm bol}$ and
285: L$_{\rm bol}$ as input parameters (Smith \cite{1998Ap&SS.261..169S,
286: 2000IrAJ...27...25S, s02}, Myers et al. \cite{1998ApJ...492..703M}) will
287: lead to uncertainties since both depend on the viewing angle, in contrast to
288: the envelope mass (Andr\'e et al. \cite{2000prpl.conf...59A}), which is
289: inferred from optically thin dust emission. Hence, the absolute values of the
290: determined ages and masses are model dependent but the relative values can be
291: used to put these youngest known protostars in an evolutionary sequence.
292: Calculated ages and final masses from all models are listed in
293: Table\,\ref{evolmod}.
294:
295: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
296: %%%%%%%%%%%%%%%%%%%%%% Discussion %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
297: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
298:
299: \section{Discussion}
300:
301: \label{discussion}
302:
303: \subsection{Source classification}
304:
305: We investigated SED's of 95 young protostellar sources. 59 of these objects
306: (60\,\%) possess sufficient observational data to accurately determine all
307: properties (T$_{\rm bol}$, L$_{\rm bol}$, M$_{\rm env}$). Out of these, 27
308: objects could be unambiguously classified as Class\,0 object, 23 as Class\,0/1,
309: and 9 as Class\,1 protostars (according to our definition). There is a
310: remaining relatively large number (36 objects; 40\,\%) where we do not have
311: sufficient information about the SED for a proper classification.
312:
313: There are two main gaps in the observational data which mean this large
314: fraction could not be classified properly. (1) Only data at wavelengths shorter
315: than the emission maximum and non or only insufficient sub-mm or millimeter
316: observations are available. Hence, only upper limits for T$_{\rm bol}$ and
317: lower limits for L$_{\rm bol}$ can be determined. (2) Only data at wavelengths
318: longer than the maximum and no observations or just upper limits at shorter
319: wavelengths are available. For these objects only $\beta$ and M$_{\rm env}$ can
320: be estimated properly.
321:
322: \subsection{Uncertainties}
323:
324: All determined object parameters are subject to different errors, due to their
325: calculation. The powerlaw index of the opacity is determined from the mean
326: slope of the SED at $\lambda$\,$>$\,400\,$\mu$m. Using fluxes taken with
327: different aperture sizes to determine $\beta$ will lead to different results.
328: Estimating $\beta$ as the mean slope ensures that for each object a 'mean'
329: aperture size for the (sub)mm data is used. Typical errors of the estimated
330: values are 0.3.
331:
332: The luminosity is determined by integrating the SED and adopting a distance. An
333: error of 20\% for the distance (a typically value for most of the objects with
334: d\,$<$\,500\,pc) leads to 40\% uncertainty for the luminosity. Also the
335: luminosity might be underestimated when the SED is not well sampled at the
336: emission peak, or when huge parts of the SED are only described by upper limits
337: (e.g. VLA\,1623 or HH\,24\,MMS). Additional uncertainties are expected in
338: cluster regions, where current far-infrared instruments do not possess
339: sufficient angular resolution to separate individual sources (e.g. SVS\,13\,B).
340: Non-detections or observations in the NIR or at millimeter wavelength have
341: almost no influence, since most energy is radiated near the emission maximum.
342: All together the determined source luminosities are uncertain by about 50\% in
343: most cases.
344:
345: For the bolometric temperature, datapoints at short wavelengths are important.
346: In the case of well sampled SEDs from the NIR to the sub-mm, T$_{\rm bol}$ can
347: be determined with about 5 to 10\,K accuracy. In the case of upper limits in
348: the NIR, the same accuracy applies as long as the limits are at least 5-6
349: orders of magnitude lower than the flux at the maximum of the emission. Note
350: that the sources classified as Class\,0/1 have typically NIR detections 4-5
351: orders of magnitude below the emission maximum, while for the Class\,0 sources
352: we find typical non-detections 5-7 orders of magnitude below the maximum. In
353: case of 'bad' NIR limits the uncertainty in T$_{\rm bol}$ can well exceed
354: 10\,K. Hence for our Class\,0 sources the temperatures are accurate to 10\,K,
355: while for the Class\,0/1 objects they are certain to 20\,K. Note that for older
356: sources (Class\,0/1 or Class\,1) viewing angle effects might become important.
357: They lead in case of edge-on sources to underestimates of T$_{\rm bol}$ and
358: L$_{\rm bol}$. Subsequently the L$_{\rm smm}$/L$_{\rm bol}$ ratio is
359: overestimated, influencing the source classification.
360:
361: The envelope mass is also subject of various sources of errors. Beside the
362: problems with measuring the fluxes in different apertures, the distance is used
363: in the determination. Further the applied dust opacities are uncertain by a
364: factor of two (e.g. Motte \& Andr\'e \cite{2001A&A...365..440M}), and the dust
365: temperatures are not very well known. Hence, the masses might suffer from
366: errors of a factor of three.
367:
368: How do these errors influence inferred ages and masses? The models of Smith
369: \cite{1998Ap&SS.261..169S, 2000IrAJ...27...25S, s02} and Myers et al.
370: \cite{1998ApJ...492..703M} use T$_{\rm bol}$ and L$_{\rm bol}$ as main input
371: parameter. In both cases the error in the measured luminosity leads to
372: uncertainties of about a factor of two for the estimated final masses. Errors
373: in T$_{\rm bol}$ will transform into age uncertainties which are about 30\%.
374: The model of Andr\'e et al. \cite{2000prpl.conf...59A} uses the envelope masses
375: as input. Hence inferred final masses are uncertain by a factor of three and
376: estimated source ages are uncertain by a factor of two. Note that these
377: uncertainties do not take into account systematic errors due to the various
378: assumptions made in the different evolutionary models.
379:
380: \subsection{Distribution of source parameters}
381:
382: To investigate the statistical properties of the obtained Class\,0 sample, we
383: restrict all the following discussions to sources classified as Class\,0 or
384: Class\,0/1 objects within 500\,pc.
385:
386: The power-law index of the opacity is evenly distributed from 0.0 to 2.0 with a
387: broad maximum between 0.5 and 1.5. For our sources T$_{\rm bol}$ ranges from 30
388: to 90\,K, where the 90\,K limit is due to our classification criterion for
389: Class\,0 sources. The number of sources gradually rises from 30 to 70\,K and
390: falls off at higher temperatures. One expects a rising (or at least constant)
391: number of sources towards higher temperatures, because with higher temperature
392: an increased age and a slower evolution in temperature is predicted by the
393: models. Hence, the T$_{\rm bol}$ distribution shows that there are missing
394: objects in our sample in this temperature range. This might be due to a
395: classification of these objects as Class\,1 protostars and reflect the fact
396: that sources with the same bolometric temperatures are not necessarily in the
397: same evolutionary state. The distribution of L$_{\rm bol}$ corresponds very
398: well with the mass function (see below), due to the assumptions in the
399: evolutionary models. Similarly the distribution of M$_{\rm env}$ shows a peak
400: at about one solar mass. At lower masses our sample certainly suffers from
401: incompleteness. The L$_{\rm smm}$/L$_{\rm bol}$ ratio shows a broad
402: distribution between 0.01 and 0.1. Only very few larger values are found
403: (IRAM\,04191, L\,1448\,NW).
404:
405: We further investigated if the source properties are correlated. There is no
406: correlation between the dust opacity $\beta$ and the parameters T$_{\rm bol}$
407: and L$_{\rm bol}$ (correlation coefficients (c.c.) -0.034 and -0.102). A weak
408: correlation (c.c. -0.412) is found between $\beta$ and M$_{\rm env}$, but this
409: is due to a few high envelope mass sources. Also T$_{\rm bol}$ is not
410: correlated with L$_{\rm bol}$ and M$_{\rm env}$ (c.c. 0.149 and -0.219). On the
411: other hand the envelope mass shows a correlation with the bolometric luminosity
412: (c.c. 0.660). This correlation indicates that objects with more massive
413: envelopes possess a higher luminosity and hence higher accretion rates.
414:
415: \subsection{The different evolutionary models}
416:
417: Due to the different model assumptions about the time evolution of the mass
418: accretion rates, all three investigated evolutionary models lead to different
419: values for ages and final star masses. The ages especially are very model
420: dependent. We investigate if there are correlations between the inferred values
421: for age and final mass for the different evolutionary models in our Class\,0
422: and Class\,0/1 sample.
423:
424: On first sight the ages differ significantly between the models. Ages from
425: Myers et al. \cite{1998ApJ...492..703M} are much larger than usually assumed
426: for Class\,0 sources (a couple of 10$^4$\,yrs). In the model of Andr\'e et al.
427: \cite{2000prpl.conf...59A} the evolution in the first 10000\,yrs is very fast,
428: and hence the inferred ages are on average very small ($<$\,10$^4$\,yrs).
429: However, there is a weak correlation between the ages inferred from the model
430: of Smith \cite{1998Ap&SS.261..169S, 2000IrAJ...27...25S, s02} and Myers et al.
431: \cite{1998ApJ...492..703M} (c.c. 0.425). Comparing the models from Andr\'e et
432: al. \cite{2000prpl.conf...59A} with Myers et al. \cite{1998ApJ...492..703M} and
433: Smith \cite{1998Ap&SS.261..169S, 2000IrAJ...27...25S, s02} we find no
434: correlation (c.c. 0.143, 0.263).
435:
436: The predicted final masses are much more important than the ages, since the
437: mass distribution should represent the observed initial mass function. Here we
438: investigate if the relative masses obtained from the different models are
439: comparable. There are no obvious differences in the inferred masses between the
440: models. The models of Myers et al. \cite{1998ApJ...492..703M} and Andr\'e et
441: al. \cite{2000prpl.conf...59A} are limited in their mass range (0.3-0.7 and
442: 0.2-3.0\,M$_\odot$, respectively). We find good correlations between masses
443: obtained from the different models: c.c.=0.676 for Smith
444: \cite{1998Ap&SS.261..169S, 2000IrAJ...27...25S, s02} and Myers et al.
445: \cite{1998ApJ...492..703M}; c.c.=0.656 for Smith \cite{1998Ap&SS.261..169S,
446: 2000IrAJ...27...25S, s02} and Andr\'e et al. \cite{2000prpl.conf...59A} and
447: c.c.=0.771 for Myers et al. \cite{1998ApJ...492..703M} and Andr\'e et al.
448: \cite{2000prpl.conf...59A}).
449:
450: \subsection{The Mass Function}
451:
452: The evolutionary scheme by Smith \cite{1998Ap&SS.261..169S,
453: 2000IrAJ...27...25S, s02} allows the comparison of all three main observational
454: quantities (T$_{\rm bol}$, L$_{\rm bol}$, M$_{\rm env}$) with the model. Also
455: there is no restriction on the final mass as in the other models. Further, the
456: assumed mass accretion rates are very similar to accretion rates obtained by
457: hydrodynamical simulations of star formation (e.g. Klessen
458: \cite{2001ApJ...550L..77K}, Schmeja \& Klessen \cite{2004A&A...419..405S}).
459: Hence it is worth to investigate the resulting final mass function obtained for
460: our sources using this model. Certainly the absolute values for the masses are
461: uncertain, but giving the good correlations to the other models (see above),
462: the relative masses, needed to determine the slope in the mass function, are
463: correct.
464:
465: In Fig.\,\ref{imf} we present the resulting mass function for our sample of
466: Class\,0 and Class\,0/1 objects within 500\,pc. The slope determined for
467: objects with M\,$>$\,0.5\,M$_\odot$ is -0.9\,$\pm$\,0.2. Considering the low
468: number of sources, and that no binary correction is done, this is in good
469: agreement with the Salpeter slope of -1.35 measured for the solar neighborhood.
470: We estimate a completeness limit in our sample for objects with final masses
471: below 0.4\,M$_\odot$.
472:
473: \begin{figure}[t]
474: \caption{\label{imf} Mass Function for the properly classified Class\,0 and
475: Class\,0/1 objects within 500\,pc, obtained from the model of Smith
476: \cite{1998Ap&SS.261..169S, 2000IrAJ...27...25S, s02}.}
477: \includegraphics[width=7.cm, height=5.9cm, bb=90 340 310 530]{f2.eps}
478: \end{figure}
479:
480: There is, however, a number of predicted very low mass sources. These are
481: objects where the model of Smith \cite{1998Ap&SS.261..169S,
482: 2000IrAJ...27...25S, s02} seems not to work properly. Looking in detail, we
483: find that the evolutionary scheme of Smith \cite{1998Ap&SS.261..169S,
484: 2000IrAJ...27...25S, s02} can explain the all observed properties (T$_{\rm
485: bol}$, L$_{\rm bol}$, M$_{\rm env}$) for most of the sources simultaneously. In
486: case of eleven objects, however, the scheme is not working (IRAS\,03256+3055,
487: IRAS\,03282+3035, B\,213, IRAM\,04191, L\,1448\,N, IRAS\,04325+2402,
488: IRAS\,04368+2557, IRAS 15398-3359, HH\,24\,MMS, NGC\,2068\,LBS\,17, VLA\,1623).
489: These sources possess a much lower luminosity (considering their T$_{\rm bol}$
490: and M$_{\rm env}$ values) than the average of the objects. There are also three
491: objects where the luminosity is much above the average (NGC\,1333-I2,
492: HH\,212-MM, L\,1641\,N). Note that this applies also for the more distant
493: objects L\,1246-SMM\,1 (luminosity lower) and Cep\,E, IRAS\,20050+2720, and
494: MonOB1\,IRAS12\,S1 (luminosity higher). For the remaining sources the model
495: explains very well all three observed properties.
496:
497: There are two reasons for this inability to explain the data of all our
498: objects. (1) One might be found in the mass accretion rates obtained by Klessen
499: \cite{2001ApJ...550L..77K} and Schmeja \& Klessen \cite{2004A&A...419..405S}.
500: It turnes out that their accretion rates on average show a time evolution
501: similar to the one used in the model of Smith \cite{1998Ap&SS.261..169S,
502: 2000IrAJ...27...25S, s02}. But on short timescales the accretion rate varies
503: significantly from this average. There are periods with lower and higher
504: accretion rates resulting in higher and lower luminosities. Hence, the very low
505: luminosity sources might represent a quiet accretion phase, while the higher
506: luminosity sources represent phases of enhanced accretion. (2) A second reason
507: might be that these sources do not accrete their mass in the same way. Hence,
508: the mass accretion rate shows a signifficantly different evolution in these
509: objects.
510:
511: In case of the latter, this suggests two different groups of sources. 'Normal'
512: sources, possessing accretion rates similar to that obtained by Schmeja \&
513: Klessen \cite{2004A&A...419..405S}, and 'abnormal' sources showing on average
514: lower luminosities/accretion rates, considering their T$_{\rm bol}$ and M$_{\rm
515: env}$ values. Note that viewing angle effects could be responsible for some of
516: these objects. These low luminosity sources might represent objects where
517: ambipolar diffusion dominates the accretion process instead of turbulence. A
518: detailed analysis and discussion of this subject can be found in Froebrich et
519: al. \cite{2004A&A.inprep.F}.
520:
521: \subsection{The different Star Forming Regions}
522:
523: Does the star formation process, in particular the time dependence of the mass
524: accretion rate vary with the star forming region or is it a uniform function?
525: Due to the low number of sources a proper statistical analysis of this question
526: is not possible yet. However, in order to see possible trends we chose four
527: regions (Perseus, Taurus, Orion, and Serpens) where we have a sufficient number
528: of confirmed Class\,0 or Class\,0/1 objects (13, 4, 10, 6, respectively) and
529: compared these regions with the average of all sources.
530:
531: In our sample of Class\,0 and Class\,0/1 sources about 25\% do not follow the
532: T$_{\rm bol}$-L$_{\rm bol}$-M$_{\rm env}$ relation of the majority of our
533: objects. These sources possess much lower luminosities than suggested by their
534: bolometric temperature and envelope mass. Hence, they might be in a phase of
535: lower mass accretion or are gaining their mass in a different way. Perseus
536: reflects the overall average with three out of 13 objects. The same applies for
537: Orion (two out of ten), where we might miss low luminosity objects due to the
538: larger distance. In Serpens all of the six sources show a 'normal' behaviour.
539: This is, according to the small number of objects, still in agreement with the
540: average. Taurus, however, shows a very different picture. Here all four
541: Class\,0 or Class\,0/1 objects fall into the low luminosity category. Objects
542: showing a 'normal' behaviour should have been easily detected since their
543: higher luminosity. Hence, even if there are just four sources this strongly
544: suggests that in a region of isolated star formation (as in Taurus) the time
545: evolution of the mass accretion rates is different from regions where stars
546: form in clusters (in agreement with previous findings from Henriksen et al.
547: \cite{1997A&A...323..549H}). Due to the bad statistics nothing can be said
548: about differences among the clusters.
549:
550: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
551: %%%%%%%%%%%%%%%%%%%%%% Conclusions %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
552: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
553:
554: \section{Conclusions}
555:
556: A literature search for photometric broadband observations of 95 confirmed or
557: candidate Class\,0 sources was conducted. To our best knowledge all available
558: broadband photometric data was used to construct SEDs from 1\,$\mu$m to
559: 3.5\,mm. If possible we determined basic properties of the sources (sub-mm
560: slope of the SED, T$_{\rm bol}$, L$_{\rm bol}$, L$_{\rm smm}$/L$_{\rm bol}$,
561: and M$_{\rm env}$). For 59 objects sufficient enough data are available for a
562: proper determination of the source parameters. 27 of these are classified as
563: Class\,0, 23 as Class\,0/1, and 9 as Class\,1 protostars.
564:
565: To investigate the statistical properties of the obtained sample of very young
566: protostars we used all objetcs within 500\,pc which are properly classified as
567: Class\,0 or Class\,0/1 to determine age and final star mass. Therefor we used
568: three different evolutionary models for protostars from Smith
569: \cite{1998Ap&SS.261..169S, 2000IrAJ...27...25S, s02}, Myers et al.
570: \cite{1998ApJ...492..703M}, and Andr\'e et al. \cite{2000prpl.conf...59A}.
571: Considering the uncertainties in the measured source properties the predicted
572: final star masses from the different models show a good correlation. The
573: absolute ages are, however, very model dependent.
574:
575: An investigation of the final star masses and the resulting mass function shows
576: a good agreement with the IMF for stars with M\,$>$\,0.5\,M$_\odot$. Our
577: obtained Class\,0 sample is limited to sources with final masses above
578: 0.4\,M$_\odot$. A number of objects (25\,\%) is found that possess a much lower
579: luminosity than the rest of our sample, considering their bolometric
580: temperature and envelope mass. These objects might be in a 'quiet' accretion
581: phase or the time evolution of their mass accretion rate is significantly
582: different from the majority. This might especially be the case in the Taurus
583: star forming region, where all identified Class\,0 or Class\,0/1 sources belong
584: to this group.
585:
586: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
587: %%%%%%%%%%%%%%%%%%%%%% Acknowledgements %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
588: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
589:
590: \begin{acknowledgements}
591:
592: We are grateful to M.\,Rengel for providing a preprint of her paper. We thank
593: the anonymous referee for very helpful comments and suggestions. M.D.\,Smith is
594: acknowledged for providing the code of his evolutionary model and his comments.
595: We thank M.P.\,Redman, T.P.\,Ray, and \`A.\,Gras-Vel\'azquez for their help.
596: This publication makes use of the Protostars Webpage hosted by the Dublin
597: Institute for Advanced Studies. D.\,Froebrich received financial support by the
598: Cosmo-Grid project, funded by the Program for Research in Third Level
599: Institutions under the National Development Plan and with assistance from the
600: European Regional Development Fund. This research has made use of the SIMBAD
601: database, operated at CDS, Strasbourg, France. This research has made use of
602: NASA's Astrophysics Data System. This publication makes use of data products
603: from the Two Micron All Sky Survey, which is a joint project of the University
604: of Massachusetts and the Infrared Processing and Analysis Center/California
605: Institute of Technology, funded by the National Aeronautics and Space
606: Administration and the National Science Foundation.
607:
608: \end{acknowledgements}
609:
610: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
611: %%%%%%%%%%%%%%%%%%%%%% Literature %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
612: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
613:
614: \begin{thebibliography}{}
615: \input{literature.tex}
616: \end{thebibliography}
617:
618:
619: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
620: %%%%%%%%%%%%%%%%%%%%%% Tables %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
621: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
622:
623: \input{tab1.tex}
624:
625: \input{tab2.tex}
626:
627: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
628: %%%%%%%%%%%%%%%%%%%%%% Appendix %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
629: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
630:
631: \begin{appendix}
632:
633: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
634: %%% notes to the individual sources
635: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
636: \input{tab3.tex}
637:
638: \input{tab4.tex}
639:
640:
641: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
642: %%% The Table with the fluxes
643: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
644: %\input{tab5.tex}
645:
646: \end{appendix}
647:
648:
649: \label{lastpage}
650:
651: \end{document}
652: