Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.atnf.csiro.au/people/Ray.Norris/papers/n204.pdf
Äàòà èçìåíåíèÿ: Thu Oct 7 11:12:43 2010
Äàòà èíäåêñèðîâàíèÿ: Fri Feb 28 07:55:09 2014
Êîäèðîâêà:

Ïîèñêîâûå ñëîâà: ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï ï
The Astrophysical Journal, 651:914Y932, 2006 November 10
# 2006. The American Astronomical Society. All rights reserved. Printed in U.S.A.

A MULTIWAVELENGTH STUDY OF YOUNG MASSIVE STAR-FORMING REGIONS. I. THE IONIZED GAS CONTENT
Guido Garay
Departamento de Astronomi , Universidad de Chile, Casilla 36-D, Santiago, Chile ´a

Kate J. Brooks

1

Australia Telescope National Facility, P.O. Box 76, Epping, NSW 1710, Australia

Diego Mardones
Departamento de Astronomi , Universidad de Chile, Casilla 36-D, Santiago, Chile ´a

and Ray P. Norris
Australia Telescope National Facility, P.O. Box 76, Epping, NSW 1710, Australia Received 2003 December 17; accepted 2006 July 26

ABSTRACT We present multifrequency radio continuum observations, made with ATCA, of a sample of 15 southern hemisphere luminous IRAS point sources, all with colors typical of compact H ii regions and associated with CS (2 ! 1) emission. The sources were observed at 1.4, 2.5, 4.8, and 8.6 GHz with angular resolutions of about 700 , 400 , 200 , and 100 , respectively. Radio emission was detected toward 14 objects. Nine IRAS objects are associated with single regions of ionized gas with simple morphologies; three are associated with two regions of ionized gas, each of which also shows simple morphologies, and two are associated with regions of ionized gas that have complex structure. The H ii regions with simple morphologies have linear diameters in the range 0.011Y 0.9 pc and electron densities in the range 103Y 105 cmþ3 and are excited by stars with luminosities in the range 9 ; 103 Y2 ; 105 L. We find that the H ii regions in our sample, most of which lie at the center of massive and dense molecular cores with H2 densities of $4 ; 105 cmþ3 and line widths of $3 km sþ1, are excited by stars with an output of UV photons of typically 3 ; 1048 sþ1.Under these conditions the regions of ionized gas reach pressure equilibrium with the dense molecular surroundings in only $5 ; 103 yr. We conclude that the main mechanism of confinement of the compact H ii regions in our sample is provided by the high density and large turbulent pressure of the surrounding molecular gas; therefore, their age can be much longer than their dynamical timescale. If the objects in our sample are representative of the Galactic IRAS sources with colors of compact H ii regions, then the problem with the large rate of massive star formation in the Galaxy might be solved. Subject headingg H ii regions -- radio continuum: ISM -- stars: early-type -- stars: formation s:

1. INTRODUCTION Young massive stars emit copious Lyman continuum photons that excite their surroundings, giving rise to dense and small (diameters less than 0.1 pc) regions of ionized gas. These regions, called ultracompact ( UC ) H ii regions, are characterized by having high emission measures, which makes them very bright at radio wavelengths. Hence, they act as powerful radio beacons of newly formed stars that are still embedded in their natal molecular clouds. For recent reviews on compact H ii regions see Churchwell (2002) and Kurtz (2002). Surveys at radio continuum wavelengths have shown the existence of a large number of UC H ii regions in our Galaxy ( Wood &Churchwell 1989b,hereafter WC89; Kurtz et al. 1994; Walsh et al. 1998). Using the IRAS Point Source Catalog ( PSC) and a color criterion that selects UC H ii regions, Wood & Churchwell (1989a) estimated that there are potentially $1650 UC H ii regions within the disk of our Galaxy. Assuming that the ages of UC H ii regions are equal to their dynamical ages, their small sizes would imply that they are very young objects, typically
1 Partially affiliated with Departamento de Astronom´a, Universidad de Chile, i Casilla 36-D, Santiago, Chile.

$5 ; 103 yr. The large number of UC H ii regions and their short dynamical ages pose a problem: the rate of massive star formation appears to be much greater than other indicators suggest (Wood & Churchwell 1989a; Churchwell 1990). Why are the ages of UC H ii regions not consistent with their short dynamical ages? Most likely the expansion of UC H ii regions is inhibited by some mechanism, so that their small sizes do not necessarily indicate that they are extremely young. We have started a multiwavelength study of a sample of 18 luminous Infrared Astronomical Satellite (IRAS ) sources in the southern hemisphere thought to be representative of young massive star-forming regions. The goal is to understand the physical and chemical differences between different stages of early evolution. The objects were taken from the Galaxy-wide survey of CS (2 ! 1) emission toward IRAS sources with IR colors typical of compact H ii regions ( Bronfman et al. 1996). We selected sources based primarily on the observed CS (2 ! 1) line profiles, looking for selfabsorbed lines consistent with inward or outward motions (e.g., Mardones 1998), or with extended line wings, possibly indicating the presence of bipolar outflows. In addition, the sources were required to have IRAS 100 m fluxes greater than 103 Jy and to be in the southern hemisphere ( < þ20 ). The luminosities of the IRAS sources, computed using the IRAS energy distribution and 914


YOUNG MASSIVE STAR-FORMING REGIONS. I.
TABLE 1 Inst rumental Parameters Phase Tracking Center (J2000.0) (2) 12 13 14 15 15 15 15 16 16 16 17 17 17 17 17 41 32 13 03 43 54 55 16 49 56 04 04 05 19 30 17.43 30.47 13.88 13.25 17.74 06.03 48.49 39.32 30.15 04.02 23.06 26.83 09.79 16.05 26.47 (J2000.0) (3) þ61 þ63 þ61 þ59 þ54 þ53 þ52 þ51 þ45 þ43 þ40 þ40 þ41 þ39 þ34 44 04 16 03 07 11 42 16 17 04 43 45 28 03 41 10.4 48.9 18.3 54.0 00.1 07.6 40.2 28.4 20.1 13.5 56.3 57.0 34.1 55.7 09.2 Synthesized Beam (arcsec ; arcsec) 1.5 (4) 7.8 7.7 9.6 9.4 7.2 10.6 10.9 7.7 8.9 10.3 10.1 9.9 10.8 11.0 13.1 ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; 6.0 6.2 6.3 6.3 6.4 5.9 5.8 6.6 5.9 5.8 6.1 6.0 6.2 6.1 6.0 4.2 4.1 5.1 4.9 4.0 5.9 5.8 4.4 5.2 5.8 5.7 5.6 6.0 5.9 8.0 2.4 (5) ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; 3.3 3.5 3.5 3.7 3.6 3.3 3.3 3.7 3.3 3.2 3.4 3.3 3.3 3.4 3.3 4.8 (6) 2.1 2.1 2.3 2.2 2.3 2.5 2.5 2.2 2.4 2.9 2.6 2.6 3.0 3.1 3.5 ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; 1.9 1.7 1.8 1.8 2.0 1.8 1.8 1.9 1.9 1.9 2.0 2.0 1.8 1.9 1.9 8.6 (7) 1.1 1.2 1.3 1.2 1.3 1.4 1.4 1.2 1.3 1.6 1.5 1.5 1.7 1.8 1.9 ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; 1.0 0.9 1.0 1.0 1.1 1.0 1.0 1.1 1.1 1.0 1.1 1.1 1.0 1.1 1.1 1.5 (8) 0.47 0.50 0.41 0.98 2.93 0.83 0.55 1.58 0.64 0.40 0.70 0.70 0.54 2.68 2.27 Noise (mJy beamþ1) 2.4 (9) 0.37 0.38 0.20 0.43 2.94 1.04 0.51 0.88 0.56 0.17 0.61 0.67 0.26 1.01 2.03 4.8 (10) 0.16 0.34 0.14 0.26 0.32 1.29 0.57 0.77 0.17 0.10 0.47 0.75 0.22 0.24 1.53

915

IRAS Source (1) 12383þ6128 13291þ6249 14095þ6102 14593þ5852 15394þ5358 15502þ5302 15520þ5234 16128þ5109 16458þ4512 16524þ4300 17008þ4040 17009þ4042 17016þ4124 17158þ3901 17271þ3439 .............. .............. .............. .............. .............. .............. .............. .............. .............. .............. .............. .............. .............. .............. ..............

8.6 (11) 0.23 0.36 0.12 0.38 0.17 1.87 1.03 0.91 0.37 0.083 0.59 1.08 0.50 0.30 0.86

Note.--Units of right ascension are hours, minutes, and seconds, and units of declination are degrees, arcminutes, and arcseconds.

distances derived by L. Bronfman (2006, private communication), are in the range 1 ; 104 Y 4 ; 105 L, implying that they contain at least one embedded massive star. We report here radio continuum observations, made with the Australia Telescope Compact Array (ATCA), at the wavelengths of 20, 13, 6, and 3 cm, with angular resolutions of about 700 , 400 , 200 ,and 100 , respectively, toward 15 objects in our sample. Radio continuum observations of the other three sources in our sample have been presented elsewhere (Garay et al. 2002, 2003). The lower angular resolution observations were designed to probe possible extended diffuse emission, and the higher angular resolution to determine the morphology of the individual regions of ionized gas. The observations at the four frequencies were designed to determine the spectral index of the emission. The main goals of these observations were to determine whether these IRAS sources are excited by single massive stars or by clusters of massive stars, to determine the characteristics of the ionized gas in each of the individual components, and to identify the main mechanisms that inhibit the expansion of UC H ii regions. The characteristics of the molecular and dust emission associated with the IRAS sources in our sample will be the subject of the second paper of the series. 2. OBSERVATIONS The observations were made at the ATCA2 in Australia, during 2000 May. The array was in the 6D configuration, which utilizes all six antennas and covers east-west baselines from 77 m to 5.9 km. We observed 15 fields, centered near the IRAS point sources, at the frequencies of 1.384, 2.496, 4.800, and 8.640 GHz, using a bandwidth of 128 MHz. The total integration times onsource at each frequency were typically 60 minutes, obtained from 6 minute scans taken over a wide range of hour angles to provide good (u, v)-plane coverage. These observations are insensitive to structures larger than 30 00 (4:8/ ), where is the observed frequency in GHz. Calibration sources were observed for 1 minute, before and after every on-source scan, in order to correct the amThe Australia Telescope Compact Array is funded by the Commonwealth of Australia for operation as a National Facility managed by CSIRO.
2

plitude and phase of the interferometer data for atmospheric and instrumental effects. Calibration of the bandpass was achieved from observations of 1740þ517. The flux density was calibrated by observing PKS 1934þ638 (3C 84), for which values of 14.94, 11.14, 5.83, and 2.84 Jy at 1.38, 2.50, 4.80, and 8.64 GHz, respectively, were adopted. Standard calibration and data reduction were performed using MIRIAD (Sault et al. 1995). Maps were made by Fourier transformation of the uniformly weighted interferometer data using the AIPS task MX. The observed sources are listed in Table 1. The resulting synthesized ( FWHM ) beams and the noise level achieved in the images are given in columns (4) Y (7) and (8) Y (11), respectively. 3. RESULTS 3.1. Overall Characteristics We detected radio emission toward 14 of the 15 observed IRAS sources. Maps of their radio continuum emission at the frequencies of 1.4, 2.5, 4.8, and 8.6 GHz are shown in Figures 1Y14. The smoother appearance of the 1.4 GHz maps is due to the lower angular resolution and the moderately high opacity of the radio continuum emission at this frequency. At higher frequencies the structure of the radio sources is resolved. The IRAS peak position, indicated in the maps as a plus sign, usually lies within the boundary of the radio emission at 1.4 GHz and in all cases agrees with the radio positions within its uncertainty (!3000 ). According to the brightness distribution of the radio emission, we find that nine IRAS objects are associated with single radio components showing simple morphologies, three IRAS objects are associated with two radio components, each of which also show simple morphologies, and two IRAS objects are associated with regions of ionized gas that have complex radio continuum structure. Radio continuum observations at 4.8 and 8.6 GHz of these two sources ( IRAS 15502þ5302 and IRAS 16128þ5109) ´ were previously reported by Marti -Hernandez et al. (2003). Of ´n the 15 individual H ii regions with simple radio structure, 9 have cometary-like morphologies, characterized by the presence of acompact bright headandadiffuselow brightness extendedtail; 3 have spherical morphologies; 1 has anirregular/core-halo structure;


TABLE 2 Observed Parameters o f R adio Emission Peak Position (J2000.0) (2) (J2000.0) (3) Simple Sources 12383þ6128 ...................................... 12 12 12 12 13 13 13 13 14 14 14 14 15 15 15 15 15 15 15 15 15 15 16 16 16 16 16 16 16 16 16 16 16 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 41 41 41 41 32 32 32 32 13 13 13 13 03 03 03 03 55 55 55 55 55 55 49 49 49 49 56 56 56 56 56 56 56 04 04 04 04 04 04 05 05 05 05 19 19 19 19 30 30 30 30 17.57 17.57 17.57 17.57 31.11 31.11 31.11 31.11 14.25 14.25 14.25 14.25 13.64 13.64 13.64 13.64 48.38 48.38 48.38 48.38 48.67 48.67 30.02 30.02 30.02 30.02 02.85 02.85 02.85 03.47 03.47 03.47 03.47 20.49 20.49 28.06 28.06 28.06 28.06 11.20 11.20 11.20 11.20 15.06 15.06 15.06 15.06 26.12 26.12 26.12 27.21 þ61 þ61 þ61 þ61 þ63 þ63 þ63 þ63 þ61 þ61 þ61 þ61 þ59 þ59 þ59 þ59 þ52 þ52 þ52 þ52 þ52 þ52 þ45 þ45 þ45 þ45 þ43 þ43 þ43 þ43 þ43 þ43 þ43 þ40 þ40 þ40 þ40 þ40 þ40 þ41 þ41 þ41 þ41 þ39 þ39 þ39 þ39 þ34 þ34 þ34 þ34 44 44 44 44 05 05 05 05 16 16 16 16 04 04 04 04 43 43 43 43 43 43 17 17 17 17 04 04 04 04 04 04 04 44 44 46 46 46 46 29 29 29 29 04 04 04 04 41 41 41 41 40.8 40.8 40.8 40.8 19.2 19.2 19.2 19.2 48.2 48.2 48.2 48.2 29.9 29.9 29.9 29.9 07.23 07.23 07.23 07.23 06.4 06.4 44.5 44.5 44.5 44.5 42.6 42.6 42.6 41.0 41.0 41.0 41.0 33.8 33.8 24.6 24.6 24.6 24.6 06.7 06.7 06.7 06.7 37.0 37.0 37.0 37.0 45.2 45.2 45.2 48.8 1.38 2.50 4.80 8.64 1.38 2.50 4.80 8.64 1.38 2.50 4.80 8.64 1.38 2.50 4.80 8.64 1.38 2.50 4.80 8.64 4.80 8.64 1.38 2.50 4.80 8.64 1.38 2.50 4.80 1.38 2.50 4.80 8.64 1.38 2.50 1.38 2.50 4.80 8.64 1.38 2.50 4.80 8.64 1.38 2.50 4.80 8.64 2.50 4.80 8.64 1.38 105. 149. 120. 79. 708. 846. 616. 273. 105. 109. 103. 56. 461. 552. 511. 224. 410. 715. 1058. 767. 122. 229. 31. 72. 134. 150. 10.5 8.1 6.0 5.2 8.2 8.1 5.6 503. 164. 979. 1577. 2274. 2029. 25. 61. 129. 184. 236. 170. 63. 41. 100. 222. 346. 4203.
a

Radio Source (1)

Frequency (GHz) (4)

Flux Density (mJy) (5)

Angular Size (arcsec) (6)

13291þ6249 ......................................

4.5 2.6 2.0 10.1 7.7 6.7 4.7 5.0 5.0 4.2 8.1 7.4 6.3 4.1 6.8 5.2 4.3 3.3

; ; ; ; ; ; ;
a

2.3 2.0 1.5 7.4 6.1 4.5 2.4 4.4 3.8 2.6 6.0 6.6 5.2 2.9 4.4 4.6 3.5 1.8

14095þ6102 ......................................

14593þ5852 ......................................

15520þ5234A....................................

; ; ; ; ; ; ; ; ; ; ;
a

15520þ5234B .................................... 16458þ4512 ......................................

0.7 ; 0.5
a a

16524þ4300A.................................... 16524þ4300B ....................................

1.7 1.3 5.4 5.0 5.4

; ; ; ; ;
a a

1.1 0.6 3.3 2.2 2.4

17008þ4040 ...................................... 17009þ4042 ......................................

0.8 0.8 19.0 10.5 10.0 7.4 5.8 2.9

; ; ; ; ; ; ; ;
a a

0.5 0.7 14.3 9.2 7.4 5.7 4.3 2.0

17016þ4124 ......................................

17158þ3901 ......................................

1.2 1.0 17.0 12.6 12.8 9.9

; ; ; ; ; ;
a

0.7 0.4 9.3 7.8 5.7 5.7

17271þ3439A.................................... 17271þ3439B ....................................

1.8 ; 0.5 1.7 ; 1.3 47.1 ; 35.5b

Complex Sources 15502þ5302A.................................... 15 15 15 15 54 54 54 54 06.39 06.39 06.39 06.39 þ53 þ53 þ53 þ53 11 11 11 11 40.1 40.1 40.1 40.1 1.38 2.50 4.80 8.64 394. 603. 1406. 1740. 6.9 3.8 3.3 2.6 ; ; ; ; 4.9 2.5 2.2 1.6


YOUNG MASSIVE STAR-FORMING REGIONS. I.
TABLE 2-- Continued Peak Position (J2000.0) (2) 15 15 15 15 15 15 15 15 15 15 15 16 16 16 16 16 16 16 54 54 54 54 54 54 54 54 54 54 54 16 16 16 16 16 16 16 05.29 05.29 05.29 05.29 06.59 06.59 07.14 09.37 09.37 10.13 10.13 38.17 38.06 40.39 40.39 40.39 40.39 43.44 (J2000.0) (3) þ53 þ53 þ53 þ53 þ53 þ53 þ53 þ53 þ53 þ53 þ53 þ51 þ51 þ51 þ51 þ51 þ51 þ51 11 11 11 11 11 11 11 12 12 11 11 16 17 17 17 17 17 17 39.5 39.5 39.5 39.5 59.4 59.4 19.3 0.3 0.3 39.3 39.3 43.2 02.9 09.2 09.2 09.2 09.2 07.8 Frequency (GHz) (4) 1.38 2.50 4.80 8.64 1.38 2.50 2.50 1.38 2.50 1.38 2.50 1.38 1.38 1.38 2.50 4.80 8.64 1.38 Flux Density (mJy) (5) 93. 68. 143. 40. 98. 59. 20. 153. 83. 127. 50. 86. 53. 2383. 2875. 2138. 570. 290. Angular Size (arcsec) (6)
a

917

Radio Source (1) 15502þ5302B ....................................

3.9 ; 1.7 2.8 ; 2.2 4.8 ; 2.1
a

15502þ5302C .................................... 15502þ5302D.................................... 15502þ5302E .................................... 15502þ5302F .................................... 16128þ5109a ..................................... 16128þ5109b .................................... 16128þ5109c .....................................

6.9 ; 3.5
a

14.4 12.0 14.7 6.5 18.7 15.0 19.7 16.1 16.1

16128þ5109d ....................................

; ; ; ; ; ; ; ; ; .. 20.5 ;

4.1 3.6 7.8 3.6 7.7 10.9 11.2 7.5 5.0 . 11.0

Note.--Units of right ascension are hours, minutes, and seconds, and units of declination are degrees, arcminutes, and arcseconds. a Unresolved. b Value from Condon et al. (1998).

1has a bipolar structure; and 1has a shell morphology.As noted by WC89, the morphological classification of compact H ii regions is not always straightforward but depends on resolution, sensitivity, and subjective judgment. Sources with cometary morphology were further subdivided into two categories: fan-shaped (7) and bow (2). The fan-shaped morphology is used to describe a cometary region that shows a bright arclike head with no limb brightening and a low brightness tail that flares away from the axis of symmetry, whereas the bow morphology is used to describe a region that shows a sharp head with a clear limb brightening of the leading edge and a tapering of the brightness away from the symmetry axis. 3.2. Individual Sources In this section we discuss the characteristics of the radio emission toward each of the IRAS sources, individually. The position, flux densities, and angular size of the detected radio sources are given in Table 2. Radio continuum spectra of the sources with the highest emission measures are presented in Figure 15. Eight of the 15 IRAS sources were previously observed by Walsh et al. (1998) at 6.7 and 8.6 GHz with angular resolution of $1B5. The morphology and peak flux density at 8.6 GHz of the radio sources reported by Walsh et al. (1998) are in good agreement with those observed by us. IRAS 12383þ6128 (G301.731+1.104).--The radio emission from IRAS 12383þ6128 arises from a single source that exhibits afan-shaped morphology, with ahead toward the northwest and a tail trailing to the southeast (see Fig. 1). The tail is resolved out at 8.6 GHz. IRAS 13291þ6249 (G307.560þ0.586).--The radio emission from IRAS 13291þ6249 arises from a compact central object surrounded by an extended halo (see Fig. 2). At 1.4 GHz the halo has major and minor axes of 10B1and 7B4, respectively, and a total flux density of 0.71 Jy. At 8.6 GHz the emission from the extended halo is resolved out. The 8.6 GHz map shows that the

compact central region has a cometary morphology, exhibiting a head with a sharp edge and a long tail trailing to the southwest and a flux density of 0.27 Jy. IRAS 14095þ6102 (G312.596+0.048).--The radio emission associated with IRAS 14095þ6102 arises from a single source. At high frequencies it exhibits a fan-shaped morphology, with a compact head and a tail trailing to the northwest (see Fig. 3). The emission from the extended tail is partially resolved at 8.6 GHz. IRAS 14593þ5852 (G319.163þ0.419).--The morphology of the radio emission from IRAS 14593þ5852 shows an irregular structure (see Fig. 4). The emission at 1.4 GHz arises from an extended structure with major and minor axes of 8B1 and 6B0, respectively, and a flux density of 0.46 Jy. At 8.6 GHz the emission arises mainly from a compact central region with major and minor axes of 4B1and2B9, respectively, and a flux density of 0.22 Jy, the extended emission being partially resolved out. IRAS 15394þ5358 (G326.474+0.697).--No radio emission was detected in any the four observed frequencies within a region of 20 centered on IRAS 15394þ5358. We establish 3 upper limits on the flux density at 4.8 and 8.6 GHz of 1.0 and 0.5 mJy, respectively. At the frequencies of 1.4 and 2.5 GHz we detected radio emission associated with IRAS 15408þ5356 and IRAS 15384þ5349, which are located $12A8eastand 12A6northwest of IRAS 15394þ5358, respectively. IRAS 15502þ5302 (G328.307+0.432).--The radio emission toward IRAS 15502þ5302 arises from a complex region of ionized gas (see Fig. 5). The region appears considerably different at low frequencies than at high frequencies. The 1.4 and 2.5 GHz maps show that the emission arisesfrom a bright source ( labeled A; peak brightness temperature, Tb , of 3600 K at 2.5 GHz), a less bright source located west of A ( labeled B), and at least four low brightness components (Tb $ 250 K at 2.5 GHz) located eastward of A ( labeled C, D, E, and F ). At 4.8 GHz we detected emission only from sources A and B, and at 8.6 GHz only source A was detected. Our maps at these two frequencies are in good agreement


Fig. 1.--ATCA maps of the radio continuum emission from IRAS 12383þ 6128. FWHM beams are shown in the lower left corner. Top:1.4 GHz map. Contour levels are (þ0:5; 0:5; 1; 2; 3; 4; 5; 6; 7; 8; 9; 10) ; 7:0 mJy beamþ1. Second:2.5 GHz map. Contour levels are ( þ 0:5; 0:5; 1; 2; 3; 5; 7; 9; 12; 15) ; 5:0mJy beamþ1. Third:4.8 GHz map. Contour levels are (þ0:5; 0:5; 1; 2; 3; 5; 7; 10; 15; 20; 25; 35; 45) ; 1:0 mJy beamþ1. Bottom: 8.6 GHz map. Contour levels are (þ0:5; 0:5; 1; 2; 3; 5; 7; 9; 12; 15) ; 1:25 mJy beamþ1.

Fig. 2.--ATCA maps of the radio continuum emission from IRAS 13291þ 6249. FWHM beams are shown in the lower left corner. Top:1.4 GHz map. Contour levels are (þ1; 1; 2:5; 5; 10; 20; 30; 40; 50; 60; 70; 80) ; 3:0 mJy beamþ1. Second: 2.5 GHz map. Contour levels are (þ1; 1; 2:5; 5; 10; 20; 30; 40; 50; 60; 70; 80; 90) ; 2:0 mJy beamþ1. Third: 4.8 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 10; 20; 30; 40; 60; 80) ; 1:0mJy beamþ1. Bottom:8.6 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9; 12; 15; 20) ; 1:5 mJy beamþ1.

918


Fig. 3.--ATCA maps of the radio continuum emission from IRAS 14095þ 6102. FWHM beams are shown in the lower left corner. Top: 1.4 GHz map. Contour levels are (þ1; 1; 3; 5; 10; 20; 30; 40; 50; 60) ; 1:2 mJy beamþ1. Second: 2.5 GHz map. Contour levels are (þ1; 1; 3; 5; 10; 20; 30; 40; 50; 60; 70) ; 0:7 mJy beamþ1. Third: 4.8 GHz map. Contour levels are (þ1; 1; 3; 5; 7; 10; 15; 20; 30; 40) ; 0:4mJy beamþ1. Bottom: 8.6 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9; 12; 15; 18) ; 0:3 mJy beamþ1.

Fig. 4.--ATCA maps of the radio continuum emission from IRAS 14593þ 5852. FWHM beams are shown in the lower left corner. Top: 1.4 GHz map. Contour levels are (þ1; 1; 3; 5; 10; 20; 30; 50; 70; 90) ; 2:5 mJy beamþ1. Second: 2.5 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9; 12; 15; 20; 30; 40; 50; 60; 70) ; 2:0mJy beamþ1. Third: 4.8 GHz map. Contour levels are (þ1; 1; 3; 5; 7; 10; 15; 20; 25; 30; 40; 50; 60) ; 1:0 mJy beamþ1. Bottom: 8.6 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9; 12; 15; 18; 21) ; 1:0 mJy beamþ1.

919


920

GARAY ET AL. ´ with those presented by Marti -Hernandez et al. (2003). The ´n emission from the four low brightness components is completely resolved out at 4.8 and 8.6 GHz. The radio emission from source A exhibits unusual characteristics. Its flux density steadily increases with frequency, whereas its angular size decreases with frequency. This is shown in Figure 16 below, which plots the flux density S and size s versus frequency in the range from 1.4 to 8.6 GHz. Power-law fits to the dependencies with frequency (dotted lines) give S / 0:9 ô 0:1 and s / þ0:5 ô 0:1 . These dependencies suggest that the electron density in the region is not uniform but decreases with the distance to the central exciting object (Olnon 1975; Panagia & Felli 1975). See further discussion in x 4.2.1. IRAS 15520þ5234 (G328.808+0.632).--The radio emission from IRAS 15520þ5234 arises from two regions of ionized gas: (1) a western component ( labeled A; see Fig. 6) exhibiting a cometary-like structure with a head with a sharp edge toward the north and a tail trailing to the south, and (2) an eastern compact component ( labeled B). At 1.4 and 2.5 GHz most of the emission arises from component A. At 8.6 GHz both sources have similar peak brightness, but the emission from the tail of component A is partially resolved out (see also Ellingsen et al. 2005). The radio continuum spectrum of component B, shown in Figure 15, is well fitted by that of a uniform density H ii region with an emission measure of 2:2 ; 108 pc cmþ6 and an angular size of 0B83. This region of ionized gas is still optically thick at 8.6 GHz. IRAS 16128þ5109 (G332.153þ0.445).--The radio emission toward IRAS 16128þ5109 has an irregular multiple-peak structure (see Fig. 7). The emission appears considerably different in the low-frequency maps than in the high-frequency maps. At 1.4 GHz the emission shows a bright (Tb $ 5500 K ), central component ( labeled c) and three low brightness temperature (Tb $ 500 K ) components ( labeled a, b, and d), located within a region of $4000 in radius surrounding the bright component. Lowercase labels have been used to avoid confusion with the notation of ´ Marti -Hernandez et al. (2003). The total flux density measured ´n in the 1.4 GHz map is 3.1 Jy. At 8.6 GHz, the emission from the outer components seen at 1.4 GHz is completely resolved out, whereas that from the central component is partially resolved out. The total flux density measured at 8.6 GHz is 0.57 Jy. The maps at 1.4, 2.5, and 4.8 GHz show that component c has a fanshaped morphology. At 1.4 GHz we measured major and minor axes of 19B7 and 11B2, respectively, and P:A: ¼ 12 . The emission from component c at 8.6 GHz shows a clumpy structure, consisting of four, and possibly five, UC components with sizes of $100 Y200 ($0.02Y0.04 pc) located within a region of $1700 ($0.3 pc) in diameter. The issue of whether these small-scale structures correspond to partially ionized globules or each of them is excited by an internal energy source is discussed in x 4.2. IRAS 16458þ4512 (G340.248þ0.373).--The radio emission from IRAS 16458þ4512 (see Fig. 8) arises from a single source that is unresolved at 1.4 and 2.5 GHz. At 8.6 GHz the source shows a fan-shaped morphology, with major and minor axes of 1B3and 0B63. The radio continuum spectrum (see Fig. 15) is well fitted by that of a uniform density region of ionized gas with an emission measure of 4:1 ; 107 pc cmþ6 and an angular diameter of 1B35. IRAS 16524þ4300 (G342.704+0.130).--The radio emission from IRAS 16524þ4300 (see Fig. 9) arises from two components: (1) a western component ( labeled A), bright at 1.4 GHz but resolved out at 8.6 GHz, and (2) an eastern component ( labeled B), bright at the higher frequencies. Component A has a

Fig. 5.--ATCA maps of the radio continuum emission from IRAS 15502þ 5302. FWHM beams are shown in the lower left corner. Top: 1.4 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9; 12; 15; 18; 22; 26) ; 8:0 mJy beamþ1. Second: 2.5 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 10; 15; 20; 30; 40; 50) ; 7:0 mJy beamþ1. Third: 4.8 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 10; 15; 20; 30; 50; 70; 90) ; 5:0 mJy beamþ1. Bottom: 8.6 GHz map. Contour levels are (þ1; 1; 3; 5; 10; 20; 30; 40; 50; 60) ; 6:0 mJy beamþ1.


Fig. 6.--ATCA maps of the radio continuum emission from IRAS 15520þ 5234. FWHM beams are shown in the lower left corner. Top: 1.4 GHz map. Contour levels are (þ1; 1; 3; 5; 10; 20; 30; 40; 50; 70) ; 3:0 mJy beamþ1. Second: 2.5 GHz map. Contour levels are (þ1; 1; 3; 5; 10; 20; 30; 40; 60; 80; 100) ; 3:0 mJy beamþ1. Third: 4.8 GHz map. Contour levels are (þ1; 1; 3; 5; 10; 15; 20; 25; 30; 35; 40; 45; 50) ; 5:0mJy beamþ1. Bottom:8.6 GHz map. Contour levels are (þ1; 1; 3; 5; 10; 15; 20; 25; 30; 35; 40) ; 4:0 mJy beamþ1.

Fig. 7.--ATCA maps of the radio continuum emission from IRAS 16128þ 5109. FWHM beams are shown in the lower left corner. Top:1.4 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9; 12; 15; 20; 30; 40; 50; 60; 70) ; 6:0 mJy beamþ1. Second: 2.5 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 10; 20; 35; 50; 65; 80; 95; 110) ; 3:0 mJy beamþ1. Third: 4.8 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 10; 15; 20; 30; 40) ; 4:0mJy beamþ1. Bottom:8.6 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9) ; 4:0 mJy beamþ1.


Fig. 8.--ATCA maps of the radio continuum emission from IRAS 16458þ 4512. FWHM beams are shown in the lower left corner. Top: 1.4 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9; 12; 15) ; 2:0 mJy beamþ1. Second: 2.5 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9; 12; 15; 18; 22) ; 2:5 mJy beamþ1. Third: 4.8 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 10; 20; 30; 40; 50) ; 1:5 mJy beamþ1. Bottom: 8.6 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 10; 20; 30; 40; 50) ; 1:5 mJy beamþ1.

Fig. 9.--ATCA maps of the radio continuum emission from IRAS 16524þ 4300. FWHM beams are shown in the lower left corner. The IRAS peak position is located $1500 west of component A. Top: 1.4 GHz map. Contour levels are (þ1; 1; 2; 3; 4; 5; 6; 7; 8) ; 0:9mJy beamþ1. Second: 2.5 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9; 11; 13) ; 0:6mJy beamþ1. Third:4.8 GHz map. Contour levels are ( þ 1; 1; 2; 3; 5; 7; 10; 15; 20) ; 0:35 mJy beamþ1. Bottom: 8.6 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9; 12) ; 0:3 mJy beamþ1.


YOUNG MASSIVE STAR-FORMING REGIONS. I.

923

Fig. 10.--ATCA maps of the radio continuum emission from IRAS 17008þ 4040. FWHM beams are shown in the lower left corner. Top: 1.4 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9; 12; 15; 20) ; 4:0 mJy beamþ1. Bottom: 2.5 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9; 12) ; 2:0 mJy beamþ1.

fan-shaped morphology, with major and minor axes at 1.4 GHz of 5B4 and 3B3, respectively. Component B has roughly a spherical morphology, with major and minor axes at 8.6 GHz of 0B76 and 0B69, respectively. IRAS 17008þ4040 (G345.499 +0.354).--The radio emission from IRAS 17008þ4040 arises from a single extended H ii region, with a fan-shaped morphology (see Fig. 10). We measured major and minor axes at 1.4 GHz of 19B0and 14B3, respectively. The source was not detected at 4.8 and 8.6 GHz, presumably because it was resolved out at these frequencies. IRAS 17009þ4042 (G345.490+0.311).--The radio emission from IRAS 17009þ4042 arises from a single source (see Fig. 11) and exhibits a fan-shaped morphology. The deconvolved angular size of the emission decreases with frequency from $8B6 at 1.4 GHz to $2B4at 8.6GHz.We find that the brightness temperature at the peak of the emission is similar at all frequencies, with a value of $5000 K, indicating that in this line of sight the ionized gas is optically thick at all frequencies. The observed characteristics suggest that the electron density toward this region is not uniform but has strong gradients (see x 4.1.2). IRAS 17016þ4124 (G345.001þ0.220).--The radio emission from IRAS 17016þ4124 arises from a single source that is unresolved at 1.4 and 2.5 GHz (see Fig. 12). The 4.8 and 8.6 GHz maps show that the source is elongated in the north-south direction, and we classify its morphology as bipolar. The radio continuum spectrum (shown in Fig. 15) is better fitted by a model of anebula with aGaussian density distribution (dashed line) than with a uniform density model (dotted line). This suggests that the

Fig. 11.--ATCA maps of the radio continuum emission from IRAS 17009þ 4042. FWHM beams are shown in the lower left corner. Top: 1.4 GHz map. Contour levels are (þ1; 1; 2:5; 5; 10; 20; 30; 40; 50; 60; 70; 80; 90) ; 4:0 mJy beamþ1. Second: 2.5 GHz map. Contour levels are (þ1; 1; 2:5; 5; 10; 20; 30; 40; 50; 65; 80; 95; 110) ; 4:0 mJy beamþ1. Third: 4.8 GHz map. Contour levels are (þ1; 1; 2:5; 5; 10; 20; 30; 40; 50; 65; 80; 95) ; 4:0mJybeamþ1. Bottom: 8.6 GHz map. Contour levels are (þ1; 1; 2:5; 5; 10; 20; 30; 45; 60; 75) ; 3:5 mJy beamþ1.


Fig. 12.--ATCA maps of the radio continuum emission from IRAS 17016þ 4124. FWHM beams are shown in the lower left corner. The IRAS peak position is located $1500 west of the radio source. Top: 1.4 GHz map. Contour levels are (þ1; 1; 2; 3; 4; 5; 7; 9; 12; 15) ; 1:5mJy beamþ1. Second:2.5 GHz map. Contour levels are (þ1; 1; 3; 5; 10; 15; 20; 30; 40; 50; 60; 70) ; 0:8 mJy beamþ1. Third: 4.8 GHz map. Contour levels are (þ1; 1; 3; 5; 10; 20; 30; 50; 70; 90; 120) ; 0:8 mJy beamþ1. Bottom: 8.6 GHz map. Contour levels are (þ1; 1; 3; 5; 10; 20; 30; 50; 70) ; 1:7 mJy beamþ1.

Fig. 13.--ATCA maps of the radio continuum emission from IRAS 17158þ 3901. FWHM beams are shown in the lower left corner. Top:1.4 GHz map. Contour levels are (þ1; 1; 2; 3; 4; 5; 6; 7; 8; 9) ; 7:0 mJy beamþ1. Second: 2.5 GHz map. Contour levels are (þ1; 1; 2; 3; 4; 5; 6; 7; 8; 9; 10) ; 3:0mJy beamþ1. Third: 4.8 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9; 11) ; 0:6 mJy beamþ1. Bottom:8.6 GHz map. Contour levels are (þ1; 1; 2; 3; 4) ; 0:5 mJy beamþ1.


YOUNG MASSIVE STAR-FORMING REGIONS. I.

925

Fig. 15.--Radio continuum spectra of compact H ii regions with high emission measures. Dotted lines: Fits to the observed spectra with a theoretical model of a homogeneous region of ionized gas. The fitted parameters, s and EM (in pc cmþ6), are given in the lower right corner. The dashed line in the bottom left panel corresponds to theoretical spectra of a region of ionized gas with a Gaussian density distribution. The triangles in the bottom right panel correspond to data from Becker et al. (1994).

Fig. 14.--ATCA maps of the radio continuum emission from IRAS 17271þ 3439. FWHM beams are shown in the lower left corner. Top: 1.4 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9; 12; 15; 20; 30; 40) ; 7:0 mJy beamþ1. Second: 2.5 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9; 12; 15; 18) ; 6:0 mJy beamþ1. Third: 4.8 GHz map. Contour levels are (þ1; 1; 2; 3; 5; 7; 9; 12; 15; 20; 25) ; 6:0mJy beamþ1. Bottom: 8.6 GHz map. Contour levels are (þ1; 1; 3; 5; 10; 20; 30) ; 4:0 mJy beamþ1.

density in this region is not homogeneous but gradually decreases outward. IRAS 17158þ3901 (G348.534þ0.973).--The radio emission from IRAS 17158þ3901 (see Fig. 13) arises from a single extended H ii region, with a fan-shaped morphology. We measure major and minor axes of 1700 and 9B3, respectively, at 1.4 GHz. The source is optically thin at all frequencies and highly resolved at 8.6 GHz. At 1.4 GHz we measure a total flux density of 0.24 Jy, whereas at 8.6 GHz we measure 0.041 Jy. The NRAO3 VLA Sky Survey ( NVSS; Condon et al. 1998) gives for this region major and minor axes of 33B3and24B8, respectively, and a flux density of 0.52 Jy. This value is 2.2 times larger than that measured by us with angular resolution of 11 00 ; 6B1, suggesting that we are resolving out emission from the most extended structures. IRAS 17271þ3439 (G353.410þ0.367).--The radio emission toward IRAS 17271þ3439 (see Fig. 14) arises from two components: (1) an extended H ii region with a shell-like morphology ( labeled B), and (2) a bright compact source ( labeled A), with an angular size of $1B5, located near the western edge of the extended structure. At 1.4 GHz the emission arises mainly from the shell component, which is associated with the IRAS source. The shell, which is not uniform but broken to the southeast, has an angular diameter at 1.4 GHz of $4500 and a flux density of 4.25 Jy.
3 The National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc.


926

GARAY ET AL.
TABLE 3 Derived Parameters D Frequency Diameter EM (kpc) (GHz) (pc) (pc cmþ6) (5) (2) (3) (4) Optically Thin H ii Regionsa 12383þ6128 ....... 13291þ6249 ....... 14095þ6102 ....... 14593þ5852 ....... 15502þ5302B ..... 15502þ5302C ..... 15502þ5302D..... 15502þ5302E ..... 15502þ5302F ..... 15520þ5234A..... 16128þ5109c ...... 16524þ4300A..... 16524þ4300B ..... 17008þ4040 ....... 17009þ4042 ....... 17158þ3901 ....... 17271þ3439B ..... 4.4 2.8 5.7 11.5 5.6 5.6 5.6 5.6 5.6 2.9 3.7 3.6 3.6 2.0 2.1 2.0 4.5 2.5 2.5 2.5 2.5 4.8 1.4 2.5 1.4 1.4 4.8 2.5 1.4 4.8 1.4 4.8 1.4 1.4 0.068 0.093 0.130 0.392 0.068 0.132 0.060 0.210 0.290 0.054 0.197 0.074 0.011 0.160 0.047 0.122 0.892 6.0 7.3 2.0 4.5 1.0 1.6 1.6 9.8 4.3 3.1 9.7 2.3 9.4 7.1 3.8 5.7 1.5 ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; 106 106 106 106 107 106 106 105 105 107 106 105 106 105 107 105 106 7.8 7.3 3.3 2.8 1.0 2.9 4.2 1.8 1.0 2.0 5.8 1.5 2.4 1.7 2.3 1.8 1.1

Vol. 651

At the high frequencies the emission arises mainly from the compact source, with the emission from the shell source being completely resolved out. Interferometric observations toward IRAS 17271þ3439 have been previously made with angular resolutions ranging from 1B5 to 4500 .The NVSS database (4500 angular resolution; Condon et al. 1998) gives for the extended source a flux density at 1.4 GHz of 6.45 Jy and major and minor angular sizes ( FWHM ) of 47B1and 35B5, respectively. The flux density measured by us is about 70% of the NVSS value, suggesting that we are resolving out emission from the most extended structures. From observations at 5 GHz with 400 angular resolution, Becker et al. (1994) reported the presence of three radio continuum sources within the region shown in Figure 14 (their peak positions are marked by the triangles in the top panel). One of these is coincident with the UC H ii region. The other two sources lie projected at peaks of the shell-like extended H ii region. We suggest that these two radio objects do not contain energy sources of their own but mark the denser parts of a clumpy shell. Finally, Walsh et al. (1998) observed the region at the frequencies of 6.7 and 8.6 GHz with angular resolutions of $1B5, detecting only the compact H ii region. The reported peak flux density at 8.6 GHz of 200 mJy beamþ1 is in good agreement with our results. The radio continuum spectrum of the emission from the compact component (shown in Fig. 15) is well fitted by the theoretical spectrum of a homogeneous, constant density, H ii region. Assuming an electron temperature of 104 K, the fit to the spectrum gives an angular size of 1B2 and an emission measure of 1:3 ; 108 pc cmþ6. 4. DISCUSSION 4.1. Parameters of the Regions of Ionized Gas The physical parameters of most of the regions of ionized gas were determined assuming that they have uniform electron density and an electron temperature of 104 K, using the formalism of Mezger & Henderson (1967). We used kinematic distances provided by L. Bronfman (2006, private communication) determined using a rotation curve with an orbital velocity of 220 km sþ1 at 8.5 kpc from the Galactic center ( Brand 1986). Table 3 lists the distance (col. [2]), observing frequency used in the calculations (col. [3]), and the derived parameters: diameter (col. [4]), emission measure (col. [5]), electron density (col. [6]), and continuum optical depth (col. [7]). An implicit requisite of this formulation is that the nebula should be optically thin at the frequency used in the calculations. This condition is valid for most of the H ii regions in our sample, except for those with emission measures greater than 7 ; 107 pc cmþ6 that have turnover frequencies !5 GHz. The physical parameters of the H ii regions with emission measure !5 ; 107 pc cmþ6 were determined by fitting the observed spectra with the theoretical spectrum of a homogeneous, constant density plasma. The fitted parameters and derived parameters are listed in Table 3.
4.1.1. Sizes and Densities

Source ATCA (1)

ne (cmþ3) (6)

(7)

; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ;

103 103 103 103 104 103 103 103 103 104 103 103 104 103 104 103 103

0.29 0.35 0.10 0.22 0.12 0.26 0.08 0.16 0.07 0.37 0.47 0.04 0.11 0.12 0.45 0.09 0.24

Optically Thick H ii Regionsb 15520þ5234B ..... 16458þ4512 ....... 17016þ4124 ....... 17271þ3439A.....
a b

2.9 3.8 2.7 4.5

.. .. .. ..

. . . .

0.012 0.025 0.016 0.026

2.2 4.1 6.3 1.3

; ; ; ;

108 107 107 108

1.1 3.3 5.2 5.8

; ; ; ;

105 104 104 104

.. .. .. ..

. . . .

Parameters derived following Mezger & Henderson (1967). Parameters derived from a fit to the spectrum using a uniform density model.

types (or ionization characteristics) of the stars exciting H ii regions is independent of the initial conditions of the surrounding medium, one expects to find that on average ne / l þ1:5 .Possible explanations for the shallower slopes are as follows: (1) UC H ii regions are excited by stars with lower luminosities (thus lower number of ionizing photons) than those exciting more extended H ii regions (Garay et al. 1993); (2) UC H ii regions are formed in the denser parts of molecular clouds with hierarchical density structure, which results in proportionally smaller regions of ionized gas ( Kim & Koo 2001).
4.1.2. Density Gradients

The regions of ionized gas in our sample have diameters l in the range 0.01Y0.9 pc and average electron densities ne in the range 103Y105 cmþ3, corresponding to compact and UC H ii regions in the classification scheme of Kurtz (2002). As already shown in several works (see review by Churchwell 2002 and references therein), there is a good correlation between the size and the density of H ii regions. For the regions in our sample, a least-squares fit to the function ne ¼ a log l × b gives a ¼ þ1:1 ô 0:1and b ¼ 2:6 ô 0:2. If the distribution of the spectral

The electron densities given in Table 3 were computed assuming a uniform density distribution. This is a rough simplification, since molecular clouds in which massive stars form are known to be inhomogeneous and show a great deal of structure. In fact, about half of the 15 H ii regions with simple brightness distributions have fan-shaped morphologies. This morphology is best explained by models in which the ionized gas is expanding in a medium with strong density gradients ( Tenorio-Tagle 1979; Bodenheimer et al. 1979). The ionized gas is radiation bounded toward the high-density medium, giving rise to the head structure, and density bounded toward the low-density medium, giving rise to the diffuse tail. The large number of fan-shaped compact H ii regions in our sample suggests that the massive cores in which they are embedded are not uniform in density and therefore that the ionized gas is flowing in an anisotropic medium. Thus, the values given in Table 3 are only representative of the average density. Auseful diagnostic of a region of ionized gas with strong density gradients is provided by the spectral index of its radio


No. 2, 2006

YOUNG MASSIVE STAR-FORMING REGIONS. I.

927

1990) or the presence of a photoevaporated disk wind ( Hollenbach et al. 1994; Yorke & Welz 1996). The dashed line in the top panel of Figure 16 corresponds to a fit to the spectrum with a model of a region of ionized gas with a power-law electron density distribution of the form ne ( r ) ¼ n
0

r 2:4 0 ; r

Fig. 16.--Flux density (top) and deconvolved angular size (bottom)as a function of frequency for IRAS 15502þ5302 component A. The dotted lines are powerlaw fits to the data. The dashed line in the top panel corresponds to theoretical spectra of a region of ionized gas with a power-law density distribution.

where n0 and r0 are the electron density and radius, respectively, at the inner boundary of the H ii region (see expressions in Olnon 1975). Since we only observe the optically thick part of the spectrum, it is not possible to determine from the fit individual values of n0 and r0 , but only a constraint between these two quantities. Assuming a distance of 5.6 kpc and Te ¼ 104 K, we obtain that 2 n0 r0 :4 ¼ 9:5 ô 0:2 pc2.4 cmþ3. There are two other H ii regions in our sample that show clear dependencies of angular size with frequency: IRAS 13291þ6249 (s / þ0:50 ô 0:08 ) and IRAS 17009þ4042 (s / þ0:66 ô 0:13 ). In these relations s corresponds to the geometric mean of the major and minor angular sizes. Both of these H ii regions have fan-shaped morphologies, suggesting that they are undergoing champagne flows and therefore are embedded in a medium with strong density gradients. The gradient in density is therefore responsible for the dependence of observed size with frequency. Kim & Koo (2001) have already suggested that the ``compact envelopeY ultracompact core'' structure seen toward several UC H ii regions (Garay et al. 1993; Kurtz et al. 1999; Kim & Koo 2001) can be understood in terms of a champagne flow combined with the clumpy structure of the massive molecular cores. 4.2. The Nature of the Complex H ii Regions The radio emission from IRAS 15502þ5302 and IRAS 16128þ 5109 shows complex morphology, exhibiting multiple components. The question arises as to whether a single star maintains the ionization of the entire H ii complex or each of the components is excited by an individual star, implying the presence of a cluster of OB stars. In the single-star view it is implicitly assumed that the region of ionized gas immediately surrounding the star is density bounded so that photons can escape and ionize the outer components and /or envelope. The number of ionizing photons per second reaching the outer components is smaller than that emitted by the central star by the factor (/4)eþ , where is the solid angle subtended by the outer component from the ionizing star and is the optical thickness from the exciting star to the distance of the outer component. In what follows we discuss each of the complex regions individually. The total far-infrared ( FIR) luminosity of IRAS 15502þ5302, computed using the IRAS fluxes (see Casoli et al. 1986), is $8:6 ; 105 L (assuming a distance of 5.6 kpc). This luminosity can be supplied by a single zero-age main-sequence ( ZAMS) star with a spectral type between O4 and O5. The rate of ionizing photons expected to be emitted by such a star is 5:4 ; 1049 sþ1 (Panagia 1973). At 1.4 GHz we identified at least six radio continuum components within a diameter of 6000 (or 1.6 pc). The most luminous of these is component A; hence, it is likely to contain the most luminous exciting star within the complex. Since component A exhibits a rising spectrum between 1.4 and 8.6 GHz, it is not straightforward to estimate the rate of UV photons required to excite this H ii region. From the observed flux density at 8.6 GHz we estimate a lower limit of 5:2 ; 1048 sþ1. Using the
4.2.1. IRAS 15502þ5302 Complex

continuum emission (Olnon 1975; Panagia & Felli 1975). For a plasma with a power-law density ne / rþq, is equal to (2q þ 3:1)/(q þ 0:5) (Olnon 1975). In addition, in the presence of density gradients, the distance from the central star at which an H ii region becomes optically thin is different at different frequencies; therefore, a dependence of the observed size with frequency is expected. We are not aware of expressions given in the literature for this dependence for H ii regions with power-law density stratifications. For collimated outflows Reynolds (1986) derived a size dependence with frequency of s / þ0:7/ ,where is the powerlaw index that describes the dependence of the jet half-width with the distance to the jet origin. A power-law spectral slope of the radio continuum emission from a compact H ii region does not, however, necessarily imply a power-law dependence of electron density with radius, but it could be the result of the presence of hierarchically clumped structures, as proposed by Ignace & Churchwell (2004). However, in this case one does not expect to observe a strong dependence of angular size with frequency. One of the most striking cases in our sample is that of radio component A in IRAS 15502þ5302, for which the flux density steadily increases with frequency and the angular size decreases with frequency (see Fig. 16). These characteristics suggest the presence of a steep gradient in the electron density. The observed dependence of the flux density with frequency (S / 0:9 ô 0:1 ) implies that the density decreases outward with radius, r, as ne / rþ2:4 . A similar density stratification has been derived toward the UC H ii region G9.62+0.19E ( Franco et al. 2000). Gradients in the electron density may indicate either the existence of density gradients in the molecular ambient cloud ( Franco et al.


928

GARAY ET AL.

Vol. 651

2 constraint derived from the spectrum (n0 r0:4 ¼ 9:5 ô 0:2 pc2.4 þ3 cm ), we find that the rate of ionizing photons needed to excite component A is Ni ¼ 1:9 ; 1049 (0:01 pc/r0 )1:8 sþ1 , where r0 is the inner radius of the H ii region. This radius is not known and difficult to estimate. For r0 ¼ 0:005 pc (see discussion below), we obtain Ni ¼ 7 ; 1049 sþ1. The rate of ionizing photons required to excite component B is 4:0 ; 1047 sþ1, equivalent to that of an O9.5 ZAMS star. The rates of ionizing photons required to excite the low brightness components C, D, E, and F are, respectively, 2:4 ; 1047 , 5:0 ; 1046 , 3:8 ; 1047 , and 3:2 ; 1047 sþ1. If ionized internally, these rates imply that these components are excited by B0.5YB0 ZAMS stars. The solid angles subtended by sources C, D, E, and F from object A are, respectively, 0.14, 0.10, 0.15, and 0.18 sr. Assuming that the optical depth from the exciting star (or stars) of component A to the distance of components C, D, E, and F is $1, we estimate that the fraction of the flux of ionizing photons emitted by the exciting star arriving at the location of the outer components is typically 4 ; 10þ3. The ionizing photons needed to explain the radio emission from the low brightness components could then be supplied if component A has an output rate of UV photons of $6 ; 1049 sþ1. This could be the case if the inner radius of the H ii region of A is $5 ; 10þ3 pc. Although the single exciting star hypothesis cannot be ruled out, the most likely explanation is that the morphology of the radio emission from IRAS 15502-5302 is due to the presence of a cluster of massive OB stars. The morphology of the 1.4 GHz radio emission is notably similar to that of the Br emission presented by Bik (2004; see his Fig. 2.12). In the Br image the low brightness radio sources detected at 1.4 GHz are clearly seen as distinct sources. In addition, Bik (2004) found that the whole nebulosity is centered on an embedded young stellar cluster. We conclude that all of the radio components contain energy sources of their own. Even the individual radio component A is likely to be excited by a group of ionizing stars ( Bik 2004).

Fig. 17.--Spectral energy distribution of IRAS 15394þ5358. Flux symbols are stars for IRAS,squares for MSX, and a circle for 1.2 mm SIMBA. The longdashed line is a fit to the spectrum using three modified blackbody functions of the form B (Td )f1 þ exp ½þ( /0 ) g, with different temperatures. The shortdashed line indicates the fit for the colder temperature component ( fit parameters indicated on the upper left).

4.2.2. IRAS 16128þ5109 Complex

The total FIR luminosity of this complex, estimated using the IRAS fluxes, is $2:4 ; 105 L (assuming a distance of 3.7 kpc). If a single star provides this luminosity, it would correspond to an O6 ZAMS star, which emits an output rate of ionizing photons of 1:2 ; 1049 sþ1 (Panagia 1973). Support for the presence of such a l uminou s obje ct i s pro vided by sin gle-dish measu r ements of the flux density at radio wavelengths that give values of 7.9 Jy at 5.0 GHz ( Haynes et al. 1979) and 9.9 Jy at 8.9 GHz ( McGee et al. 1979), implying that the flux of ionizing photons required to excite the whole H ii region is $1 ; 1049 sþ1. With ATCA we measured a total flux density from the region of 3.1 Jy at 1.4 GHz, and smaller values at higher frequencies, indicating that we resolved out a large fraction of the extended emission. Assuming the single exciting star hypothesis and that the star is located near the center of the H ii complex (component c), the solid angles subtended by components a, b, and d are 0.43, 0.25, and 0.58 sr, respectively, implying a supply of ionizing photons from the central star of 3:4 ; 1047 ,2:0 ; 1047 ,and4:6 ; 1047 sþ1.These photon rates are of the order of (slightly larger than) those required to explain the radio emission from the outer components; thus, it is possible that the ionization of the whole region is provided by a single luminous star. The 8.6 GHz observations show the presence of several very small scale structures ($0.02 pc in radius) of ionized gas embedded in the much larger structure seen at 1.4 GHz. Their densities and emission measures are typically 2 ; 104 cmþ3 and 1 ; 107 pc cmþ6, respectively. Although these structures have sizes

within the range of the hypercompact H ii regions as classified by Kurtz (2002), their emission measures are considerably smaller and are unlikely to have energy sources of their own. The smallscale structures in the central region of IRAS 16128þ5109 might be partially ionized dense globules within the large H ii region. The existence of clumps near recently formed stars is expected due to the fragmentation process of the accreting molecular gas during the gravitational collapse to form the central exciting star. We suggest that the irregular, multiple-peaked structure of the ionized gas at large and small scales observed toward IRAS 16128þ 5109 indicates the presence of large density fluctuations of the ambient gas. The lack of detection of radio emission from IRAS 15394þ 5358 (3 upper limit of 0.5 mJy) raises the question about its nature. Since the value of the luminosity is key to the possible interpretations of this result, we compiled the spectral energy distribution of IRAS 15394þ5358 using all known flux densities (see Fig. 17). The data consist of the IRAS fluxes, mid-infrared (8Y21 m) fluxes taken from the Midcourse Space Experiment (MSX )database ( Price 1995), and the 1.2 mmflux measured by us using SIMBA /SEST ( Paper II ). Integrating under the fitted curve, and assuming that this object is at a distance of 2.8 kpc, we obtain a bolometric luminosity of 1:1 ; 104 L. This luminosity corresponds to that of a B0.5 ZAMS star, which emits a flux of UV photons of 1:7 ; 1046 sþ1 (Panagia 1973). If embedded in a uniform density medium, we expect that this star would produce an H ii region with a flux density of $25 mJy at optically thin radio frequencies. Possible explanations for the lack of radio emission 4.3. IRAS 15394þ5358: A Luminous Radio-Silent Source


No. 2, 2006

YOUNG MASSIVE STAR-FORMING REGIONS. I.
TABLE 4 Characteristics of E xciting S tars Radioa Nc0 47 þ1 LUV (104 L) (4) 2.6 3.5 2.8 13.9 2.4 2.1 3.9 2.2 1.8 4.0 2.3 0.94 0.90 3.7 22.1 LFIR (104 L) (5) 4.6 2.7 7.2 55.4 4.8 7.0 8.2 6.4 2.0 21.5c .. . 4.3c .. . 37.5c .. . Far-Infrared
b

929

Source (1) 12383þ6128 .................... 13291þ6249 .................... 14095þ6102 .................... 14593þ5852 .................... 16458þ4512 .................... 17016þ4124 .................... 17009þ4042 .................... 17008þ4040 .................... 17158þ3901 .................... 15520þ5234A.................. 15520þ5234B .................. 16524þ4300A.................. 16524þ4300B .................. 17271þ3439A.................. 17271þ3439B ..................
a Since dust within observations are lower b Spectral type and limits. c Luminosity of the

(10

s) (2)

Spectral Type (3) B0 O9.5 B0 O6.5 B0 B0 O9.5 B0 B0 O9.5 B0 B0.5 B0.5 O9.5 O6

Spectral Type (6) O9 B0 O8 O5 O9 O8 O7.5 O8 B0 O6 ... O9 ... O5.5 ...

ö Nc (1047 sþ1) (7)

2.4 5.6 3.0 61.4 2.1 1.3 7.3 1.6 0.75 8.0 1.8 0.11 0.095 6.7 104

12. 2.8 26. 333. 13. 25. 32. 22. 1.2 101. ... 10. ... 211. ...

the H ii region absorbs UV radiation, the spectral type and luminosity of the ionizing star derived from the radio limits. rate of ionizing photons are derived assuming that a single star provides the entire luminosity. These are upper whole region, including the two radio components.

from IRAS 15394þ5358 are that (1) it contains a group of nonionizing stars, which account for the total bolometric luminosity but produce a negligible amount of ionizing photons; and (2) it corresponds to a young luminous massive protostar undergoing accretion at a high rate, thus quenching the development of an H ii region. With the currently available observational data it is not possible to discern between these two alternatives. 4.4. Characteristics of the Exciting Stars The determination of the characteristics of the exciting star of compact H ii regions is not a straightforward task. As discussed below, lower and upper limits on the spectral type of the ionizing star can be derived from radio continuum and from FIR observations, respectively. Table 4 summarizes the properties of the exciting stars of compact H ii regions with simple morphologies derived from both radio continuum and FIR observations. From the radio continuum flux density at an optically thin frequency one can determine the rate of ionizing photons needed to excite the H ii region. From this rate one can then infer the spectral type of the exciting star. However, since a fraction of the Lyman continuum photons could be absorbed by dust within the H ii regions, this is actually a lower limit. Column (2) of Table 4 gives the rate of ionizing photons, Nc0 , required to maintain the ionization of each of the individual radio sources, and columns (3) and (4) give, respectively, the spectral type and luminosity, LUV ,of asingle ZAMS star that provides that photonrate( Panagia 1973). The compact regions of ionized gas in our sample are surrounded by dense and warm regions of circumstellar gas ( Paper II ). The dust in these regions absorbs all of the stellar radiation, either directly or after being processed in the nebula, reemitting the absorbed energy in the FIR. The FIR luminosity provides then an important constraint on the characteristics of the ionizing source. Column (5) of Table 4 gives the FIR luminosity of the region, LFIR,derived from the IRAS fluxes following the prescription of Casoli et al. (1986) using the distances given in column (2) of Table 3. Columns (6) and (7) give, respectively, the spectral type

and the total rate of ionizing photons, Ncö , calculated assuming that a single ZAMS star provides the entire infrared luminosity. Due to the coarse angular resolution of the IRAS observations, the FIR luminosity may include, in addition to the contribution from the stars exciting the UC H ii region, the contribution from less massive stars in their neighborhood that are not hot enough to contribute to ionization. Therefore, the values obtained assuming that the whole IR luminosity is produced by a single ionizing star are actually upper limits. For the H ii regions with simple morphologies, the range in the luminosity of the exciting star spanned by the upper and lower limits is narrow, with the FIR luminosities (LFIR) being larger than the luminosities derived from the radio continuum observations (LUV) by typically a factor of 2. As mentioned before, it is not easy to disentangle which of the two possible explanations for the difference in luminosities, namely, the presence of dust within the region of ionized gas or the contribution to the FIR luminosity from a group of stars that are not hot enough to contribute to ionization, dominates. Assuming that absorption of UV radiation by dust within the H ii region is the dominant effect, we derive that the fraction of Lyman continuum photons absorbed by dust, fd ¼ ö 1 þ (Nc0 /Nc ), is typically 0.8, suggesting that in compact H ii regions a significant amount of UV photons are absorbed by dust within the ionized gas. For the reasons given above this value of fd should be taken as an upper limit. We find that the objects in our sample are considerably less luminous than the sources in the sample of WC89. The later sample consists of 75 UC H ii regions, selected mainly from the Galactic plane survey of Wink et al. (1982), with strong FIR emission. Due to the selection criteria, these regions, which were used for defining the age problem, are probably representative of the most luminous massive star-forming regions ( MSFRs) in the Galactic plane. Our sample consists of 18 IRAS sources with colors of UC H ii regions and CS emission, chosen from the list of Bronfman et al. (1996), and with line profiles indicative of infall and /or outflow motions. Due to these spectral line characteristics,


930

GARAY ET AL.

Vol. 651

Fig. 18.--FIR luminosity distribution of sources in our sample (top) and the Wood & Churchwell (1989a) sample (bottom).

winds from luminous O stars that are in motion relative to the ambient molecular cloud (van Buren et al. 1990), as well as massive primordial disks surrounding newly formed OB stars that are photoevaporated by the UV photons from the star ( Hollenbach et al. 1994). Since bow shocks are static configurations, neither expanding nor contracting, UC H ii regions would live as long as the moving star remains embedded within the molecular cloud. In the photoevaporated disk model, the reservoir of dense gas within the disk may last for a million years or more, producing a UC H ii region that could live much longer than indicated by their dynamical ages because they are constantly being replenished by adense circumstellar reservoir. Eventhoughthese two mechanisms have been observationally identified as producing small regions of ionized gas (van Buren et al. 1990; Hollenbach et al. 1994), the fraction of compact H ii regions with these morphologies is small. Only two of the cometary sources in our sample have detailed characteristics that can be best explained in terms of bow shocks, and only one could be possibly interpreted as a photoevaporated disk ( IRAS 17016þ4124). Thus, another mechanism that could explain the confinement of UC H ii regions needs to be identified. There are two results from our multiwavelength studies that provide key information to understanding why UC H ii regions remain compact for a period of time much longer than indicated by a simple dynamical age calculation (td ¼ R/cs , where R is the size and cs the sound speed of the ionized gas). First, most of the compact H ii regions are projected toward the center of massive cores that have molecular hydrogen densities of typically 4 ; 105 cmþ3 and line widths in optically thin molecular lines of typically 3 km sþ1 (Paper II ). This suggests that the compact H ii regions are deeply embedded in massive cores, being surrounded by molecular ambient gas with high densities and large turbulent motions. Second, most of the compact H ii regions are excited by embedded stars emitting an output of ionizing photons of 3 ; 1048 sþ1. The dense and turbulent ambient medium and the low number of ionizing photons result in very small H ii region equilibrium radii ( De Pree et al. 1995). Using the condition of final pressure equilibrium, 2nf kTe ¼ n0 mH àv2 ,and the relation between the initial and final Stromgren radius (see Garay & Lizano 1999), the ¨ equilibrium radius can be written as 1=3 2 Nu 4 ; 105 cmþ3 Rf ¼ 0:034 3 ; 1048 sþ1 n0 2=3 4=3 Te 3 km sþ1 ; pc; 4K 10 àv
=3

our sample is thought to be representative of young MSFRs. The difference in luminosities is illustrated in Figure 18, which shows histograms of the distribution in the FIR luminosities for the two samples. The average FIR luminosity and Lyman conö tinuum photon rate (Nc ) of the objects in the WC89 sample, excluding complex sources, are, respectively, 7:0 ; 105 L and 2:5 ; 1049 photons sþ1 (data taken from their Table 18). For the single objects in our sample the geometric average of the FIR luminosity and Lyman continuum photon rate are 8 ; 104 L and 1:6 ; 1048 photons sþ1, respectively. Since both samples satisfy the same IRAS color criterion, the question arises as to which is more representative of the characteristics of the exciting stars of Galactic IRAS sources with colors of compact H ii regions. In this regard we investigated possible differences in the IRAS properties between the two samples that could provide some insight on this issue. We found that there is no difference between the two samples with respect to average IRAS colors. A Student's t-test shows that both samples are likely to be drawn from the same parent distribution (the null hypothesis) with a confidence level of 95%. 4.5. The Confinement Mechanism of UC H ii Regions Mechanisms that have been proposed to explain the confinement of UC H ii regions are bow shocks supported by the stellar

Ï 1÷

where n0 is the density of the ambient gas, Nu is the rate of ionizing photons emitted by the exciting star, and àv is the line width of the molecular emission from the ambient gas. The time to reach pressure equilibrium is teq ¼ 4:8 ; 103 1=3 Nu 4 ; 105 cmþ3 3 ; 1048 sþ1 n0 2=3 þ1 7=3 Te 3 km s ; yr: 4K 10 àv
2=3

Ï 2÷

For n0 ¼ 4 ; 105 cmþ3, Nu ¼ 3 ; 1048 sþ1,and àv ¼ 3km sþ1, we find that the equilibrium radius is $0.03 pc and that the time to reach pressure equilibrium is $5 ; 103 yr. We conclude that most of the UC H ii regions in our sample reach pressure equilibrium with their dense and turbulent surroundings in very short


No. 2, 2006

YOUNG MASSIVE STAR-FORMING REGIONS. I.

931

times, of $5 ; 103 yr. The identification of the dynamical timescale with the age of compact H ii regions is therefore not suitable. If our sample is more representative of the characteristics of the exciting stars of Galactic IRAS sources with colors of compact H ii regions, then the problem with the large rate of massive star formation in the Galaxy might be solved, since the age of compact H ii regions can be much longer than their dynamical timescale. The above analysis implicitly assumes a uniform density ambient medium. However, as discussed previously, the most common situation is one in which the ambient medium has density gradients, giving rise to cometary H ii regions. In this case, the dynamical expansion of the ionized gas in the head is rapidly slowed down due to the high density of the ambient gas as described above. Due to the streaming of the high-density ionized gas toward the low-density region, the recombinations in the head are, however, reduced and therefore the ionization front is not completely stalled but advances into the densest region. The velocity of the ionization front is, however, considerably smaller than the sound speed of the ionized gas (e.g., Whitworth 1979). Thus, the actual age of a champagne H ii region can be much greater than the dynamical timescale of the head structure, as has been already pointed out by Kim & Koo (2001). 5. SUMMARY We made multifrequency radio continuum observations, using ATCA, of 15 luminous IRAS point sources with colors of UC H ii regions and CS (2Y1) emission. These are thought to be massive star-forming regions in early stages of evolution. The objectives were to determine the morphologies and physical properties of the ionized gas content in these sources and to investigate the characteristics of their exciting stars. Our main results and conclusions are summarized as follows. Radio emission was detected toward 14 IRAS objects. Twelve IRAS sources are found to be associated with either one (nine objects) or two (three objects) radio sources, each of which shows a simple morphology. The most common morphology among the 15 H ii regions with simple radio structure is cometary (nine). Three have spherical morphologies, one has an irregular/core-halo structure, one has a bipolar structure, and one has a shell morphology. Two IRAS sources are found to be associated with regions of ionized gas that exhibits complex radio continuum structure. Although the single exciting star hypothesis cannot be ruled out, the high luminosity and complex morphology of one of these sources (IRAS 15502þ5302) are most likely produced by the presence of

a cluster of embedded exciting OB stars. No radio emission was detected toward IRAS 15394þ5358 to a 3 level of 0.5 mJy at 8.6 GHz. Further observations are needed to discern whether this object corresponds to a precursor of a UC H ii region or is the host of a group of nonionizing stars. The H ii regions have linear diameters in the range 0.011Y 0.9 pc and average electron densities, determined assuming uniform conditions, in the range 103Y105 cmþ3. The regions of ionized gas in our sample are, however, unlikely to be homogeneous. In fact, a large fraction of the H ii regions have fan-shaped morphologies, indicating the presence of gradients in the electron density. One of the most striking examples is radio component A toward IRAS 15502þ5302, whose radio characteristics indicate a strong stratification of the ambient medium. The observed dependence of the flux density with frequency (S / 0:9 ô 0:1 ) implies that the density decreases outward with radius, r, as ne / rþ2:4 . We find that the objects in our sample are considerably less luminous than the sources in the sample of WC89, which were used for defining the age problem. The geometric average of the FIR luminosity and Lyman continuum photon rate of the single objects in our sample, 8 ; 104 L and 1:6 ; 1048 photons sþ1, respectively, are an order of magnitude smaller than those of the corresponding objects in the WC89 sample. Most of the H ii regions in our sample are embedded within dense massive cores, which have molecular hydrogen densities of $4 ; 105 cmþ3 and turbulent line widths of $3 km sþ1. Under these conditions the regions of ionized gas reach pressure equilibrium with their dense and turbulent molecular surroundings in only $5 ; 103 yr. We conclude, based on our multifrequency observations, that the high density and large turbulent pressure of the molecular gas surrounding UC H ii regions are the main mechanisms of their confinement. If our sample is more representative of the characteristics of the exciting stars of Galactic IRAS sources with colors of compact H ii regions than the WC89 sample, then the problem with the large rate of massive star formation in the Galaxy might be solved, since the age of compact H ii regions can be much longer than their dynamical timescale.

K. B., G. G., and D. M. gratefully acknowledge support from the Chilean Centro de Astrof´sica FONDAP 15010003. G. G. i also acknowledges support from the Chilean Fondecyt Project 1010531.

REFERENCES Becker, R. H., White, R. L., Helfand, D. J., & Zoonematkermani, S. 1994, Garay, G., Rodriguez, L. F., Moran, J. M., & Churchwell, E. 1993, ApJ, 418, ´ ApJS, 91, 347 368 Bik, A. 2004, Ph.D. thesis, Amsterdam Univ. Haynes, R. F., Caswell, J. L., & Simons, L. W. J. 1979, Australian J. Phys. Bodenheimer, P., Tenorio-Tagle, G., & Yorke, H. W. 1979, ApJ, 233, 85 Astrophys. Suppl., 48, 1 Brand, J. 1986, Ph.D. thesis, Leiden Univ. Hollenbach, D., Johnstone, D., Lizano, S., & Shu, F. H. 1994, ApJ, 428, Bronfman, L., Nyman, L. 8., & May, J. 1996, A&AS, 115, 81 654 Casoli, F., Dupraz, C., Gerin, M., Combes, F., & Boulanger, F. 1986, A&A, Ignace, R., & Churchwell, E. 2004, ApJ, 610, 351 169, 281 Kim, K., & Koo, B. 2001, ApJ, 549, 979 Churchwell, E. 1990, A&A Rev., 2, 79 Kurtz, S. 2002, in ASP Conf. Ser. 267, The Earliest Stages of Massive Star ------. 2002, ARA&A, 40, 27 Birth, ed. P. A. Crowther (San Francisco: ASP), 81 Condon, J. J., Cotton, W. D., Greisen, E. W., Yin, Q. F., Perley, R. A., Taylor, Kurtz, S., Churchwell, E., & Wood, D. O. S. 1994, ApJS, 91, 659 G. B., & Broderick, J. J. 1998, AJ, 115, 1693 Kurtz, S. E., Watson, A. M., Hofner, P., & Otte, B. 1999, ApJ, 514, 232 De Pree, C. G., Rodriguez, L. F., & Goss, W. M. 1995, Rev. Mex. AA, 31, 39 ´ Mardones, D. 1998, Ph.D. thesis, Harvard Univ. ´ Ellingsen, S. P., Shabala, S. S., & Kurtz, S. E. 2005, MNRAS, 357, 1003 Marti -Hernandez, N. L., van der Hulst, J. M., & Tielens, A. G. G. M. 2003, ´n Franco, J., Kurtz, S., Hofner, P., Testi, L., Garci -Segura, G., & Martos, M. ´a A&A, 407, 957 2000, ApJ, 542, L143 McGee, R. X., Newton, L. M., & Butler, P. W. 1979, MNRAS, 189, 413 Franco, J., Tenorio-Tagle, G., & Bodenheimer, P. 1990, ApJ, 349, 126 Mezger, P. G., & Henderson, A. P. 1967, ApJ, 147, 471 Garay, G., Brooks, K., Mardones, D., & Norris, R. P. 2003, ApJ, 587, 739 Olnon, F. M. 1975, A&A, 39, 217 Garay, G., Brooks, K., Mardones, D., Norris, R. P., & Burton, M. G. 2002, ApJ, Panagia, N. 1973, AJ, 78, 929 579, 678 Panagia, N., & Felli, M. 1975, A&A, 39, 1 Garay, G., & Lizano, S. 1999, PASP, 111, 1049 Price, S. D. 1995, Space Sci. Rev., 74, 81


932

GARAY ET AL.
Walsh, A. J., Burton, M. G., Hyland, A. R., & Robinson, G. 1998, MNRAS, 301, 640 Whitworth, A. 1979, MNRAS, 186, 59 Wink, J. E., Altenhoff, W. J., & Mezger, P. G. 1982, A&A, 108, 227 Wood, D. O. S., & Churchwell, E. 1989a, ApJ, 340, 265 ------. 1989b, ApJS, 69, 831 ( WC89) Yorke, H. W., & Welz, A. 1996, A&A, 315, 555

Reynolds, S. P. 1986, ApJ, 304, 713 Sault, R. J., Teuben, P. J., & Wright, M. C. H. 1995, in ASP Conf. Ser. 77, Astronomical Data Analysis Software and Systems IV, ed. R. A. Shaw, H. E. Payne, & J. J. E. Hayes (San Francisco: ASP), 433 Tenorio-Tagle, G. 1979, A&A, 71, 59 van Buren, D., Mac Low, M.-M., Wood, D. O. S., & Churchwell, E. 1990, ApJ, 353, 570