1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass{emulateapj}
3: \usepackage{natbib,epsfig}
4:
5: \begin{document}
6:
7: \newcommand{\hii}{{\rm H}{\sc ii}}
8: \newcommand{\uchii}{{\rm UCH}{\sc ii}}
9: \newcommand{\etal}{{\it et. al.}}
10: \newcommand{\htwo}{${\rm H_2}$}
11: \newcommand{\nhone}{NH$_3$(1,1)}
12: \newcommand{\nhtwo}{NH$_3$(2,2)}
13: \newcommand{\nhthree}{NH$_3$(3,3)}
14: \newcommand{\ammonia}{NH$_3$}
15: \newcommand{\methanol}{CH$_3$OH}
16: \newcommand{\htwoo}{H$_2$O}
17: \newcommand{\hsixalpha}{H66$\alpha$}
18: \newcommand{\thcoone}{$^{13}$CO ${(J=1\rightarrow0)}$}
19: \newcommand{\ceioone}{C$^{18}$O ${(J=1\rightarrow0)}$}
20: \newcommand{\ceio}{C$^{18}$O}
21: \newcommand{\cseo}{C$^{17}$O}
22: \newcommand{\nh}{NH$_3$}
23: \newcommand{\um}{$\mu$m}
24: \newcommand{\percc}{cm$^{-3}$}
25: \newcommand{\persqcm}{cm$^{-2}$}
26: \newcommand{\emunits}{pc~cm$^{-6}$}
27: \newcommand{\kms}{km~s$^{-1}$}
28: \newcommand{\kmspc}{km~s$^{-1}$~pc$^{-1}$}
29: \newcommand{\vlsr}{${\rm v}_{\rm lsr}$}
30: \newcommand{\msun}{M$_\odot$}
31: \newcommand{\lsun}{L$_\odot$}
32: \newcommand{\jyperbeam}{Jy beam$^{-1}$}
33: \newcommand{\mjyperbeam}{mJy beam$^{-1}$}
34: \newcommand{\jyperbeamkms}{Jy beam$^{-1}$km~s$^{-1}$}
35: \newcommand{\arcseconds}{$''$}
36: \newcommand{\pv}{P-V}
37:
38: \title{Spherical Infall in G10.6-0.4: Accretion Through an Ultracompact \hii\ Region}
39:
40: \author{Peter K. Sollins\altaffilmark{1}, Qizhou Zhang\altaffilmark{1}, Eric Keto\altaffilmark{1}, Paul T. P Ho\altaffilmark{1}}
41: \altaffiltext{1}{Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA, 02138, psollins@cfa.harvard.edu}
42:
43: \begin{abstract}
44:
45: We present high resolution ($0.''12 \times 0.''079$) observations of
46: the ultracompact \hii\ region G10.6-0.4 in 23~GHz radio continuum and
47: the \nhthree\ line. Our data show that the infall in the molecular
48: material is largely spherical, and does not flatten into a molecular
49: disk at radii as small as 0.03~pc. The spherical infall in the
50: molecular gas matches in location and velocity the infall seen in the
51: ionized gas. We use a non-detection to place a stringent upper limit
52: on the mass of an expanding molecular shell associated with pressure
53: driven expansion of the \hii\ region. These data support a scenario in
54: which the molecular accretion flow passes through an ionization front
55: and becomes an ionized accretion flow onto one or more main sequence
56: stars, not the classical pressure-driven expansion scenario. In the
57: continuum emission we see evidence for externally ionized clumps of
58: molecular gas, and cavities evacuated by an outflow from the central
59: source.
60:
61: \end{abstract}
62:
63: \keywords{stars: formation --- ISM: individual (G10.6-0.4) --- \hii\, regions -- accretion}
64:
65: \section{Introduction} \label{sec:intro}
66:
67: The ultracompact (UC) \hii\ region G10.6-0.4, at a distance of 6.0~kpc
68: \citep{dow80}, is the brightest member of a complex of variously
69: compact \hii\ regions \citep{woo89b} in an area of active star
70: formation. The associated IRAS point source, IRAS 18075-1956, has a
71: luminosity of $9.2 \times 10^5$~\lsun\ \citep{cas86} and has colors
72: that meet the criteria of \citet{woo89a} for an \uchii\ region. G10.6
73: is known to be embedded in a hot molecular core (HMC)
74: \citep{bra83,ho86,plu92}. The core is thought to contain at least
75: 1200~\msun\ of gas within a radius of 0.2~pc, based on an analysis of
76: a variety of dust continuum measurements \citep{mue02}, and at least
77: 3300~\msun\ of gas within 1.1~pc based on \ceio\ and \cseo\
78: measurements \citep{hof00}. Previous studies of the inversion lines of
79: \ammonia\ have determined that the molecular core is rotating and
80: collapsing inward toward the \uchii\ region
81: \citep{ho86,ket87a,ket88,ket90}. In these studies, using the \nhone\
82: and \nhthree\ lines, rotation is seen at size scales from 1~pc down to
83: 0.08~pc, and infall is detected in the form of red-shifted absorption
84: seen against the continuum source. \methanol\ and \htwoo\ masers are
85: seen distributed linearly in plane of the rotation
86: \citep{wal98,hof96}, while OH masers seem to be distributed along the
87: axis of rotation \citep{arg00}. In \ceioone, \citet{ho94} see
88: $10^3$~\msun\ of dense (n$\sim10^6$\percc), rotating gas in a
89: flattened (0.3$\times$0.1~pc) disk-like structure. At the highest
90: resolution achieved in earlier work, infall and rotation in the
91: molecular gas were seen simultaneously in absorption, showing that the
92: molecular gas was spiraling inward on size scales comparable to the
93: size of the \uchii\ region. Our new observations resolve the \uchii\
94: region and show clearly that the molecular gas is infalling mostly
95: spherically toward the \uchii\ region, with only slow rotation and
96: little flattening in the plane of rotation.
97:
98: Recent observations of the ionized gas within the \uchii\ region
99: showed that the ionized gas is also spiraling inward toward the stars
100: at the center of the \uchii\ region \citep{ket02a}. Subsequent
101: theoretical work showed that in small \hii\ regions, the gravitational
102: effect of the central star(s) can overcome the thermal pressure of the
103: ionized gas causing the molecular accretion flow to pass through the
104: \hii\ region boundary and continue inward as an ionized accretion flow
105: \citep{ket02b}. In this model, the \hii\ region boundary exists as a
106: standing R-type ionization front within a continuous accretion
107: flow. These results differ from classical treatments of the pressure
108: driven expansion of \hii\ regions, which predict outward motion of the
109: ionized gas as soon as the \hii\ region is formed
110: \citep{str39,spitzer}. In the classical model for pressure driven
111: expansion, the \hii\ region boundary, after a very short phase as a
112: moving R-type front, will develop a characteristic double front
113: structure composed of an isothermal shock followed by a moving D-type
114: ionization front. In this model, as the \hii\ region expands, most of
115: the displaced molecular material remains between the shock and the
116: ionization front as a dense outward moving shell, which snow-plows
117: ahead of the \hii\ region. In the alternative model of \citet{ket02b},
118: however, the accretion flow passes through a standing R-type
119: ionization front at the \hii\ region boundary and continues toward the
120: star(s) as an ionized flow. In that case there will be no swept-up,
121: dense molecular layer at the boundary, and all the molecular gas will
122: be moving inward. Ionized accretion flows represent a fundamentally
123: different mode of accretion than that seen in low-mass star
124: formation. In the ionized accretion flow scenario a main sequence star
125: accretes gas which was once part of a molecular accretion flow, but
126: has been photo-ionized as it approaches the star. G10.6 contains the
127: first example of evidence for a massive accretion flow continuing
128: through an ionization front. Our new observations constitute a
129: stringent non-detection of any outward moving molecular layer thus
130: contradicting the pressure driven expansion model (see Section
131: \ref{sec:nature}).
132:
133: \section{Observations} \label{sec:obs}
134:
135: We observed the \uchii\ region G10.6 with the NRAO Very Large Array
136: (VLA)\footnote{The National Radio Astronomy Observatory is a facility
137: of the National Science Foundation operated under cooperative
138: agreement by Associated Universities, Inc.} on February 1, 2002, with
139: the phase center at $\alpha(2000)=\rm{18^h10^m28^s.68}, \,
140: \delta(2000)=-19^o 55'49''.07$. We observed the (3,3) inversion line
141: of \ammonia\ at 23.870130 GHz with 63 spectral channels of width
142: 48.828 kHz (0.61 \kms) for a total bandwidth of 3.125 MHz (38.7 \kms)
143: centered on \vlsr$=10$ \kms, and 1.3~cm continuum with a bandwidth of
144: 15.6 MHz. The array was in the A-configuration, yielding a
145: uniform-weighted synthesized beam of width $0.''12 \times 0.''072$ for
146: a physical resolution of 0.0034$\times$0.0021~pc or 700$\times$430~AU.
147:
148: We observed the quasars 3C286, 3C273 and 1733-130 for flux, bandpass
149: and phase calibration respectively. Self-calibration of the source
150: amplitudes and phases resulted in a noise level of 0.18 (0.14)
151: \mjyperbeam\ in the uniform (natural) weighted continuum map, and 1.9
152: (1.5) \mjyperbeam\ in each uniform (natural) weighted channel map,
153: about 3 times the thermal noise limit. Expressed as a temperature, our
154: sensitivity in a natural weighted continuum map is about 25~K and in a
155: natural weighted channel map is about 280~K. The physical temperature
156: of the molecular gas around G10.6 is estimated to be only 110~K at the
157: ionization front \citep{ket90}. Thus, none of the molecular gas is
158: detectable in thermal line emission in channel maps at the $3\sigma$
159: level. But since the continuum has a peak brightness temperature of
160: 6900~K, the line absorption can be detected at roughly $25\sigma$ in a
161: single channel. The quality of the self-calibration solutions and
162: improvements in the K-band receiver system at the VLA have resulted in
163: 25 times better sensitivity in our \nhthree\ channel maps than in the
164: previous best existing \nhthree\ data for this source \citep{ket88},
165: with 3 times better spatial resolution and 2 times better velocity
166: resolution.
167:
168: \section{Results} \label{sec:results}
169:
170: \subsection{The Structure of the \hii\ Region} \label{sec:cont}
171:
172: Figure \ref{fig:contmap} shows the 23~GHz continuum map in both
173: panels. The color-scale in the left panel is stretched to emphasize
174: the structure in the weaker emission while the right panel is scaled
175: to show clearly the structure in the strongest emission. Four features
176: in particular, marked A1-A6, B1-B5, C, and D in Figure
177: \ref{fig:contmap}, are worth noting.
178:
179: First, to the east of the main \uchii\ region, there are six arc
180: shaped continuum sources marked A1 through A6 in Figure
181: \ref{fig:contmap}. Several of these are arcuate, and convex, pointing
182: back toward the central \uchii\ region. All except for A1 are
183: brightest on their western edges as if they are externally ionized by
184: photons from the \hii\ region or mechanically by a wind. Arcs A2-A5
185: show line absorption (see Figure \ref{fig:mom1}) suggesting that they
186: are embedded inside the HMC. Because A1 is also bright enough to show
187: detectable absorption, but does not, A1 must either be in front of the
188: densest gas, or the gas toward A1 might have lower column density or
189: might not be warm enough to populate the (3,3) state. The upper limit
190: on the optical depth of the \nhthree\ main hyperfine line toward A1 is
191: $\tau_\nu < 0.6$. By contrast, toward the continuum peak, we detect an
192: optical depth of the main hyperfine line in excess of 65.
193:
194: Second, there are larger scale linear structures extending to the
195: south and northeast from the main \uchii\ region marked B1 and B2 in
196: Figure \ref{fig:contmap}. In the right panel of Figure
197: \ref{fig:contmap} there are smaller scale linear features marked B3,
198: B4 and B5. One can envision a wide angle outflow with an evacuated
199: cavity whose edges are defined by B1, B3 and B4 on the southwest and
200: by B2 and B5 on the northeast, parallel to the axis of rotation of the
201: molecular gas, which is about 40 degrees east of north
202: \citep{ket87b,ket88}. The outflow would fill the ``V'' shape on the
203: northeast edge of the \uchii\ region. There is no \nhthree\ absorption
204: above the sensitivity limit of optical depth 0.3 toward source B1. We
205: suggest that B1 defines the edge of the cavity evacuated by the
206: southwest side of the outflow, protruding toward the observer out of
207: the densest molecular gas. Ongoing observations in the \hsixalpha\
208: line in this region have measured the motion of ionized gas in the
209: cavities, with the southwest side (B1, B3, B4) moving toward the
210: observer, and the northwest side (B2, B5) moving away. The general
211: shape of the brightest continuum emission is extended perpendicular to
212: the rotation axis with an aspect ratio of about 2, which is consistent
213: with there being an ionized disk perpendicular to the proposed outflow
214: along the rotation axis.
215:
216: Third, the slightly resolved, almost circular peak just to the
217: northwest of the main \hii\ region, marked C in Figure
218: \ref{fig:contmap}, may be a separate ``hyper-compact'' \hii\
219: region. Because of its regular shape, it seems unlikely that the
220: ionization is caused by UV photons leaking out of the central \hii\
221: region. If this is a distinct, extremely compact \hii\ region, its
222: radius, $\sim 0''.08$ or about 500~AU, is consistent with
223: gravitational trapping of the ionized gas by 100~\msun\ of stars
224: \citep{ket02b}.
225:
226: Fourth, while there are some sharp edges in the continuum emission,
227: for instance, in sources B1 and B2, the edges of the main \uchii\
228: region itself are quite gradual, even ``frothy''. This froth is marked
229: D in Figure \ref{fig:contmap}. This frothy edge structure is present
230: all around the \uchii\ region except in the regions which we propose
231: to be the evacuated outflow cavity. So in a scheme in which accretion
232: is going on through a flattened structure and outflow is going on
233: along the rotation axis, we can associate the froth with
234: irregularities specifically in that part of the ionization front
235: through which accretion is taking place. The irregular structures
236: could be related to varying density in the gas crossing the ionization
237: front, or photo-ionization of a very clumpy circumstellar
238: environment. Several possible ionization front instabilities are
239: discussed in \citet{ket89}, but because the continuum in this part of
240: the map is weak we do not have line data to help us distinguish
241: between different possibilities.
242:
243: The much sharper ionization fronts associated with B1 and B2 could be
244: due to the much higher velocities in the outflow and an evacuated
245: environment in the direction of the outflow. Finally the isolated
246: nature of A2-A6, with no intervening ionized structures to the central
247: source, also suggests a clumpy circumstellar environment. To be
248: consistent with the frothy structure seen towards the central region,
249: the size-scale and separation of clumps must increase away from the
250: center. Future work, in which we will map the radio continuum at high
251: resolution in a variety of frequencies, will allow us to carefully
252: determine the spectral energy distribution for comparison to models,
253: such as those of \citet{ign04}, of clumpy \uchii\ regions.
254:
255: \subsection{Resolving the Infall Motions} \label{sec:mom1}
256:
257: Our absorption line data clearly show that infall dominates the
258: kinematics of the molecular gas. Figure \ref{fig:mom1} is a map of the
259: first moment (that is the flux-weighted average velocity) of the main
260: hyperfine component of the \nhthree\ absorption line. While previous
261: work done at lower angular resolution relied on \pv\ diagrams to
262: determine the velocity of the gas as a function of position
263: \citep{ket87b,ket88}, our higher resolution data allow us to map the
264: velocity of the line at every position in the two dimensional map, not
265: just along selected one dimensional cuts. The velocity of the
266: absorbing gas ranges from \vlsr$=+3.5$~\kms\ shown as white in Figure
267: \ref{fig:mom1}, to \vlsr$=-2.5$~\kms\ shown as purple. The velocity of
268: the ambient gas, as determined from widespread emission seen at lower
269: resolution and subsequent radiative transfer modeling is
270: \vlsr$=-3.0$~\kms\ \citep{ho86,ket90}. Thus the white points in Figure
271: \ref{fig:mom1}, i.e. the most red-shifted points, show the peak of the
272: infall. Moving away from that peak position in any direction, the
273: infall velocity decreases, forming concentric rings of red, orange,
274: and yellow in Figure \ref{fig:mom1}. The smooth, circular pattern of
275: radially decreasing red-shift is indicative of spherical infall, in
276: which the line-of-sight infall velocity gets projected out as one
277: moves away from the center of infall \citep{ho86}. The velocity
278: pattern is dominated by this ``bulls-eye'' pattern created by the
279: spherical infall.
280:
281: There is also evidence in Figure \ref{fig:mom1} of the rotation that
282: was seen at larger radii, out to 1~pc \citep{ho86}. Earlier work has
283: shown that the rotation axis runs from southwest to northeast, with
284: the rotating gas coming toward the observer in the southeast and going
285: away from the observer in the northwest \citep{ho86,ho94}. The
286: rotation is apparent in Figure \ref{fig:mom1} only as a perturbation
287: to the ``bulls-eye'' pattern of spherical infall. The effect of the
288: rotation is to shift the apparent center of infall along the plane of
289: rotation, toward the northwest, where the rotational velocity is away
290: from the observer along the line of sight. Although the rotation is
291: evident in Figure \ref{fig:mom1}, the circular velocity at a radius of
292: 0.03~pc is 2~\kms, which is less than the infall velocity. The
293: molecular gas is therefore spiraling inward to the \uchii\ region
294: without first settling into a disk.
295:
296: Despite the rotation seen at larger radii, we see no evidence for the
297: molecular gas settling into a disk at a radius of 0.03~pc. The optical
298: depth of the absorbing gas, calculated from the ratio of the
299: absorption in the main and satellite hyperfine components, shows an
300: overall radial drop-off from a peak in the center and a slight trend
301: toward higher optical depths in the plane of rotation, suggesting a
302: somewhat flattened, but mostly spherical density structure as opposed
303: to a geometrically thin molecular disk. By contrast, if the molecular
304: accretion flow were geometrically thin in the plane of rotation of the
305: core, as in, for instance, a rotationally supported disk, then any
306: inclination of the molecular disk to the line of sight would result in
307: high optical depth on one side of the \hii\ region and much lower
308: optical depth on the other side. This optical depth pattern is not
309: detected. The more spherical density structure is consistent with the
310: mainly spherical velocity structure of the flow. While rotation
311: dominates at larger radii as shown by \citet{ho86, ket88, ho94}, our
312: data show that at a radius of 0.03~pc, rotation does not dominate. No
313: disk is formed at that radius. While the rotational velocities
314: increase inward \citep{ket88}, the infall increases inward more
315: quickly, and dominates as the gas crosses the \uchii\ region
316: boundary. The molecular gas apparently spirals into the \uchii\ region
317: without first settling into a molecular disk. This certainly does not
318: preclude the possibility of an ionized disk inside the \uchii\ region.
319:
320: Most of the line absorption must take place very close to the \uchii\
321: region. Because we see the effects of projection in both the spherical
322: infall and the rotation components of the velocity structure, the
323: geometry requires that the absorbing gas be immediately outside the
324: \hii\ region and wrap closely around the continuum backlight. If the
325: absorbing gas were not right up against the continuum source, the
326: radial motions would not show significant projection effects away from
327: the center of infall, and the velocity would not return to the ambient
328: velocity at the edge of the absorption, as it does. For this reason,
329: we can associate the infall and rotation velocities seen in the
330: absorbing gas with the radius of the \uchii\ region. We also note that
331: the optical depth falls off by a factor of $\sim 100$ from the center
332: of the \uchii\ region to the continuum sources A1-A6, which is
333: consistent with the centrally condensed and centrally heated molecular
334: core model of \citet{ket90}. A strongly centrally condensed core
335: naturally locates most of the absorbing material, and therefore most
336: of the absorption, physically close to the continuum source.
337:
338: An important connection can be made between the infall we see in the
339: molecular gas just outside the \uchii\ region, and the motions of
340: ionized gas inside the \uchii\ region. \citet{ket02a} found that the
341: \hsixalpha\ radio recombination line showed infall motions in the
342: ionized gas inside the \uchii\ region. Figure 2 of that paper is a map
343: of the first moment of the \hsixalpha\ line at 1\arcseconds\
344: resolution. Comparing \citet{ket02a} and our Figure \ref{fig:mom1}
345: shows that the \hsixalpha\ and \nhthree\ lines have the same peak
346: infall velocity at the same position. This shows that the infall in
347: the molecular gas continues right through the ionization front and
348: into the ionized gas.
349:
350: \subsection{The Nature of High Mass Star Formation} \label{sec:nature}
351:
352: Our maps confirm the presence of simultaneous rotation and infall in
353: the molecular gas, and ongoing work in the \hsixalpha\ line has shown
354: both infall over a large area, and outflow in a more localized
355: jet-like structure in the ionized gas \citep{ket04}. The simultaneity
356: of these different phenomena with main sequence stars, as evidenced by
357: the presence of the \uchii\ region, points to a schedule for the
358: formation of massive stars that is compressed relative to the schedule
359: for low mass stars. Stellar structure calculations suggest that a
360: massive star should reach the main sequence before collapse in the
361: molecular core has finished \citep{nor00,beh01}. We find in our
362: observations evidence for this compressed evolutionary
363: sequence. Furthermore, the presence of inward motion in the ionized
364: gas and its apparent continuity with the molecular accretion flow
365: suggests that the presence of an \uchii\ region does not necessarily
366: end accretion as in the classical model for pressure driven expansion
367: of \hii\ regions. The lack of a molecular disk contrasts with the case
368: of M17, in which a 20~\msun\ star is accreting through a molecular
369: disk which extends beyond a radius of 15,000~AU \citep{chi04}. While
370: the two cores have similar angular velocities, roughly 10~\kmspc, at a
371: radius of 0.1~pc, the much larger central mass in G10.6 allows infall
372: to dominate and prevent the formation of a large molecular disk.
373:
374: The lack of any detectable blue-shifted absorption places a stringent
375: upper limit on the mass of an expanding molecular shell, which would
376: be the necessary product of pressure-driven expansion of the \uchii\
377: region. As noted in Section \ref{sec:intro}, the classical model of an
378: over-pressured \hii\ region expanding into a molecular medium, results
379: in a dense, isothermally shocked, molecular layer surrounding the
380: \hii\ region. This layer should contain nearly all of the molecular
381: gas originally in the volume now occupied by the \hii\ region. We
382: estimate that the mass of such a layer, assuming a radius for the
383: \hii\ region of 0.025~pc, an average molecular gas density over the
384: volume of the \uchii\ region of $10^6$~\percc, LTE conditions at
385: 150~K, and an \ammonia\ abundance of $10^{-7}$ relative to \htwo\
386: \citep{van98}, would be 4~\msun resulting in a predicted optical depth
387: in the main hyperfine component of $\tau \sim 2$. We detect no such
388: layer in our line data. At the peak of the continuum, our $3\sigma$
389: upper limit on the optical depth of an expanding molecular layer is
390: $\tau = 0.05$, corresponding to a shell-mass of roughly 0.1~\msun. We
391: detect no outward motion; only inward motions in both the molecular
392: and ionized gas are seen. This stringent upper limit on the mass of
393: any expanding shell, and the fact that the molecular accretion flow
394: which we observe at the 0.03~pc scale continues in the ionized gas,
395: both argue against the pressure-driven expansion model and for the
396: existence of an ionized accretion flow.
397:
398: The central mass in this region can be estimated in two ways. First,
399: we note that the infall peaks around \vlsr$=+3.5$~\kms. The systemic
400: velocity of the local gas is \vlsr$=-3.0$~\kms\ \citep{ket90}. If the
401: gas has a free-fall collapse speed of 6.5~\kms\ at a radius of
402: 0.03~pc, the central mass must be about 150~\msun. We can also
403: calculate the mass of the central star(s) by inferring the rate of
404: Lyman continuum photon production. Because the ionized gas is
405: infalling, we assume the density goes like $\rm{n(r) =
406: n_0(\frac{r_0}{r})^{(3/2)}}$ with $\rm{r_0 = 0.03}$~pc, and $n_0 =
407: 10^5$~\percc\ . Then, following \citet{ket03}, we balance ionizations
408: with recombinations as
409: \begin{displaymath}
410: \rm{N_{Ly} = \int_{R_*}^{R_{HII}} \alpha 4 \pi r^2 n^2(r) dr = 4 \pi r_0^3 n_0^2 \alpha~ln(\frac{R_{HII}}{R_*})} \nonumber
411: \end{displaymath}
412: where $\rm{N_{Ly}}$ is the number of Lyman continuum photons,
413: $\rm{R_*}$ is the radius of the star, $\rm{R_{HII}}$ is the radius of
414: the \hii\ region, and $\alpha$ is the recombination
415: coefficient. Assuming $\rm{R_*= 10 R_{\odot}}$, $\rm{R_{HII}} =
416: 0.03$~pc as above, and $\rm{\alpha= 2 \times 10^{-13}~cm^3~s^{-1}}$,
417: we get $\rm{N_i = 2.2 \times 10^{50} s^{-1}}$. This corresponds to 4
418: O4V stars with a total mass of 175 \msun, and a total luminosity of
419: $4.4 \times 10^6$~\lsun\ \citep{vac96}. This combination of stars,
420: which matches the Lyman continuum output and, roughly, the central
421: mass as derived from the free-fall velocity, is not unique. However,
422: we can be certain that at least several stars must be responsible for
423: such a large Lyman continuum emission rate.
424:
425: The complicated structure of the ionized gas suggests clumpy structure
426: in the surrounding molecular gas. This clumpy structure at very high
427: density ($> 10^6$~\percc) is consistent with the idea that this single
428: HMC may be capable of forming many stars. The central \uchii\ region
429: has formed at the center of the HMC, but must contain more than one
430: star. The overall contraction of the core has already formed multiple
431: stars, and may fragment further at smaller scales.
432:
433: \bibliographystyle{apj}
434: \bibliography{bib_entries}
435:
436: %\begin{figure*}
437: \begin{figure}
438: \epsscale{1.0}
439: %\epsscale{1.15}
440: \plotone{fig1.eps}
441: \caption{The 1.3 cm continuum map. In the left panel, the colors range
442: linearly from -1 to 8 \mjyperbeam\ although the peak in the map is 24
443: \mjyperbeam. This is to emphasize the weaker emission structures which
444: are labeled A1-A6, B1-B5, C and D in the map. In the right panel, the
445: colors range linearly from -1 to 24 \mjyperbeam, showing the entire
446: range of emission strengths in order to emphasize structures in the
447: stronger emission. The contours are -1, 1, 2, 3, 6, 9, 12, 15, 18, 21,
448: 24, \& 27 $\times$ 0.8 \mjyperbeam. The map was made with uniform
449: weighting and the synthesized beam ($0.''12 \times 0.''079$) is shown
450: in the lower right of each panel.}
451: \label{fig:contmap}
452: \end{figure}
453: %\end{figure*}
454:
455: \begin{figure}
456: \epsscale{0.6}
457: %\epsscale{1.1}
458: \plotone{fig2.eps}
459: \caption{The map of the first moment of the main hyperfine component
460: of the \nhthree\ line. The colors range from \vlsr = -3 to
461: +4~\kms. The contour, shown for reference, is the 0.8 \mjyperbeam\
462: contour from the 1.3~cm continuum map shown in Figure
463: \ref{fig:contmap}. The line data were mapped by the AIPS task IMAGR
464: with natural weighting and a u-v taper at 750~kilolambda so that the
465: resulting synthesized beam ($0.''26 \times 0.''22$) better matched the
466: size scale of velocity pattern.}
467: \label{fig:mom1}
468: \end{figure}
469:
470: \end{document}