1: \documentclass[12pt,preprint]{aastex}
2:
3: \newcommand\OUCHII{UCH\,{\sc ii}}
4: \newcommand\UCHII{UCH}
5: \newcommand\HII{H\,{\sc ii}}
6: \newcommand\CII{C\,{\sc ii}}
7: \newcommand\HI{H\,{\sc i}}
8: \newcommand\Hi{H110$\alpha$}
9: \newcommand\He{He110$\alpha$}
10: \newcommand\Ca{C110$\alpha$}
11: \newcommand\kms{km~s$^{-1}$}
12: \newcommand\cmt{cm$^{-2}$}
13: \newcommand\CeO{C$^{18}$O}
14: \newcommand\tCO{$^{12}$CO}
15: \newcommand\thCO{$^{13}$CO}
16: \newcommand\NHthree{NH$_3$}
17: \newcommand\muG{$\mu$G}
18: \newcommand\Msun{M$_{\odot}~$}
19: \newcommand\Lsun{L$_{\odot}~$}
20: \newcommand\cmthree{cm$^{-3}~$}
21: \newcommand\cmsix{cm$^{-6}~$}
22: \newcommand\dycm{dyne cm$^{-2}~$}
23: \newcommand\etal{et al.}
24: \newcommand\be{\begin{equation}}
25: \newcommand\ee{\end{equation}}
26: \newcommand\bea{\begin{eqnarray}}
27: \newcommand\eea{\end{eqnarray}}
28: \newcommand\La{Ly${\alpha}$~}
29: \newcommand\ddeg{$^{o}$}
30: \newcommand\alfven{Alf$\acute{v}$en~}
31:
32: %
33: % Define \lsim and \gsim per J. S.'s fmt:
34: %
35: \catcode`\@=11
36: \newcommand{\gsim}{${\mathrel{\mathpalette\@versim>}}$}
37: \newcommand{\lsim}{${\mathrel{\mathpalette\@versim<}}$}
38: \newcommand{\@versim}[2]{\lower 2.9truept \vbox{\baselineskip 0pt \lineskip
39: 0.5truept \ialign{$\m@th#1\hfil##\hfil$\crcr#2\crcr\sim\crcr}}}
40: \catcode`\@=12
41:
42: %\newcommand\bbeta{\boldmath$\beta$}
43:
44: \slugcomment{To appear in ApJ Letters}
45:
46: \shorttitle{Magnetic fields near \OUCHII\ regions}
47: \shortauthors{D. Anish Roshi}
48:
49: \begin{document}
50:
51: \title{Magnetic fields at the periphery of \OUCHII\ regions from carbon recombination line observations}
52:
53: \author{D. Anish Roshi\altaffilmark{1}}
54: %\affil{Raman Research Institute, Sadashivanagar, Bangalore 560 080, India}
55: %\email{anish@rri.res.in}
56: \altaffiltext{1}{Raman Research Institute, Sadashivanagar, Bangalore 560 080, India;
57: anish@rri.res.in }
58:
59:
60: \begin{abstract}
61: Several indirect evidences indicate a magnetic origin for the
62: non-thermal width of spectral lines observed toward molecular
63: clouds. In this letter, I suggest that the origin of the non-thermal width of
64: carbon recombination lines (CRLs) observed from photo-dissociation
65: regions (PDRs) near ultra-compact \HII\ regions is magnetic
66: and that the magnitude of the line width is an estimate of the \alfven speed.
67: The magnetic field strengths estimated based on this suggestion
68: compare well with those measured toward molecular clouds with
69: densities similar to PDR densities.
70: I conclude that multi-frequency CRL observations
71: have the potential to form a new tool to determine
72: the field strength near star forming regions.
73: \end{abstract}
74:
75: \keywords{
76: ISM:magnetic fields -- radio lines:ISM --
77: ISM:lines and bands -- ISM:molecules -- MHD -- \HII\ regions}
78:
79: \section{Introduction}
80: \label{sec1}
81:
82: Observations of high-density molecular line tracers toward ultra-compact
83: \HII\ regions (\UCHII s) reveal that these \HII\ regions are embedded
84: in dense (\gsim $10^5$ \cmthree) molecular clouds
85: (eg. \nocite{cetal90}Churchwell, Walmsley \& Cesaroni 1990;
86: \nocite{c02}Churchwell 2002 and references there in).
87: %The high density of the ambient medium results in the formation of a sharp
88: %boundary between the ionized and neutral regions.
89: Far ultra-violet (FUV; 6 to 13.6 eV) radiation from massive stars
90: within the \UCHII s
91: heats the dense molecular material in this interface,
92: producing a Photo Dissociation Region (PDR)
93: (see \nocite{ht97} Hollenbach \& Tielens 1997).
94: %The PDR consists of ionized carbon, since
95: %the ionization potential of carbon is $<$ 13.6 eV.
96: Carbon recombination lines (CRLs) from such
97: regions have been detected toward a large number
98: of \UCHII s, establishing the presence of
99: dense PDRs near most \UCHII s
100: (\nocite{retal05a}Roshi \etal\ 2005a).
101: %Carbon recombination lines (CRLs) from such
102: %regions have been detected toward a large number
103: %of \UCHII s, establishing the presence of
104: %dense PDRs near most \UCHII s
105: %(\nocite{retal05a}Roshi \etal\ 2005a).
106:
107: The observed width of the CRLs from PDRs associated with
108: \UCHII s is typically between 4 and 8 \kms. The gas temperature
109: in the PDR, obtained by non-LTE modeling of CRL emission at
110: multiple frequencies, is in the range 200 to 1000 K
111: (\nocite{retal05b}Roshi \etal\ 2005b;
112: \nocite{getal98}Garay \etal\ 1998;
113: \nocite{nwt94} Natta, Walmsley \& Tielens 1994).
114: Thus the expected thermal contribution to the line width
115: is at least a factor of two smaller than the observed width
116: (see Fig.~\ref{fig1}). This larger observed width indicates that
117: the CRL width is dominated by
118: supersonic motions. In fact, a similar situation exists
119: in molecular clouds. It has long been recognized that
120: the observed width of spectral lines from molecular clouds
121: have a non-thermal component and this component is often supersonic
122: (eg. \nocite{bmw64}Barrett, Meeks and Weinreb 1964). The origin
123: of the non-thermal component of the line width is now
124: attributed to either ``turbulence'' (\nocite{metal74}Morris \etal\ 1974;
125: \nocite{ze74}Zuckerman \& Evans 1974) or \alfven waves
126: (\nocite{am75}Arons \& Max 1975; \nocite{m75}Mouschovias 1975;
127: see also \nocite{sal87}Shu, Adams \& Lizano 1987 and references there in).
128:
129: Detailed theoretical models for both ``turbulence'' and
130: \alfven waves in molecular clouds are yet to be worked out.
131: But ``turbulence'' without any magnetic field
132: is a less plausible proposition since observations demand supersonic
133: turbulence and the rapid dissipation in shocks cannot
134: sustain such turbulent motions to time scales larger than the free-fall
135: time (eg. \nocite{sal87}Shu \etal\ 1987). The
136: decay of \alfven waves are slow compared to supersonic turbulence
137: though simulations show that they may also need a driving mechanism
138: (\nocite{m03}MacLow 2003).
139: Thus \alfven waves appear to be a more tenable explanation for
140: the observed line widths
141: (\nocite{am75}Arons \& Max 1975; \nocite{m75}Mouschovias 1975).
142:
143: Observational data also seem to indirectly support the magnetic origin
144: of the line width. Analysis of the width of spectral lines observed
145: toward molecular clouds shows power-law relationships between
146: the non-thermal component of the velocity dispersion, the
147: cloud size, density and the magnetic field strength
148: (\nocite{l81}Larson 1981; \nocite{th86}Troland \& Heiles 1986;
149: \nocite{mg88a}Myers \& Goodman 1988a, 1988b;
150: \nocite{mp95}Mouschovias \& Psaltis 1995; \nocite{c99}Crutcher 1999).
151: These relationships
152: are expected in self-gravitating, magnetically supported clouds
153: since the \alfven waves in such clouds affect the
154: observed line width (\nocite{m87}Mouschovias 1987; Myers \&
155: Goodman 1988a, 1998b).
156:
157: In this letter, I use the observed non-thermal widths of CRLs
158: to estimate the magnetic field strength in PDRs near UCHs.
159: The field strength is estimated by assuming that the
160: non-thermal velocity dispersion of carbon lines give an estimate of
161: the \alfven speed in the PDR. This assumption is
162: based on a comparative study of the characteristics exhibited
163: by the CRL and molecular line data. The data and the comparative
164: study are presented in Sections~\ref{sec2} and \ref{sec3}
165: respectively. The magnetic field strength
166: is obtained by combining the velocity dispersion of CRL
167: and the density derived from modeling carbon line emission
168: (see Section~\ref{sec4}).
169: To our knowledge the only attempt to compute the magnetic
170: field using CRLs was made by \nocite{v89}Vallee (1989), where
171: he had applied a model for shocks at the edge of molecular
172: clouds to deduce the field strength.
173:
174: \section{Carbon recombination line data}
175: \label{sec2}
176:
177:
178: %Far-ultraviolet (FUV) radiation from young, massive stars produce PDRs
179: %at the surface of molecular clouds. Such PDRs are often observed at the
180: %molecular cloud--\HII\ region interface (eg. \nocite{st78}Silverglate \&
181: %Terzian 1978).
182: %It is well established that the physical properties
183: %of the PDRs can be derived by modeling multi-frequency CRL emission from
184: %these regions (eg. \nocite{retal05a}Roshi \etal\ 2005a).
185: %While earlier
186: %PDR studies were mostly confined to evolved \HII\ regions,
187: %\nocite{retal05b}Roshi \etal\ (2005b) establish that dense PDRs are
188: %common near \UCHII s.
189: For the present analysis I use CRL data toward
190: \UCHII s since the PDR thickness is small in these cases and therefore
191: large scale motions (such as outflows) may not affect the line
192: width (see Section~\ref{sec5}). Table.~\ref{tab1} summarizes the
193: CRL data and results obtained from modeling this data toward 14 \UCHII s.
194: Listed are the source name,
195: observed CRL transition, FWHM of CRLs\footnote{The relationship
196: $\Delta V = \sqrt{8ln(2)}~\sigma_{NT}$ is used in this paper to relate the
197: FWHM line width $\Delta V$ to the dispersion $\sigma_{NT}$ for a Gaussian
198: line model}, PDR gas temperature, carbon ion density,
199: number density of hydrogen atoms in the PDR, line-of-sight extent of
200: the PDR ($L_{\parallel}$) and the estimated magnetic field
201: strength (see Section~\ref{sec4}).
202:
203: The quality of the CRL data obtained toward all 14 sources is good;
204: the uncertainties are mostly in the parameters derived by modeling
205: line emission.
206: The frequencies of the selected CRLs are below $\sim$ 15 GHz. As
207: inferred from modeling, carbon line emission is dominated by
208: stimulated emission at these frequencies (\nocite{nwt94} Natta \etal\ 1994;
209: \nocite{retal05a}Roshi \etal\ 2005a).
210: The data toward the first 7 sources in Table~\ref{tab1} were obtained
211: using interferometric observations.
212: The line widths for the first six sources were obtained from
213: profiles averaged over regions where lines were detected. The
214: properties of the PDR toward W3A were not well constrained; I
215: have taken representative model parameters from
216: \nocite{kag98}Kantharia, Anantharamaiah \& Goss (1998).
217: Data toward the remaining 7 sources were obtained with the Arecibo telescope.
218: Modeling of CRL emission toward these sources was limited by:
219: (a)poor angular resolution, (b)flux density calibration error and
220: (c) detection of line only near 9 GHz.
221: Thus modeling this data set has given only
222: constraints on the physical properties of the PDR.
223: The PDR gas temperature of 500 K for the Arecibo sources
224: and W48A, W49G and W49J listed in Table~\ref{tab1} is a representative value.
225: The carbon ion density and $L_{\parallel}$
226: obtained from modeling depend non-linearly on the gas temperature.
227: Converting the ion density to neutral density is somewhat uncertain
228: due to the unknown depletion factor and the fraction of molecular
229: hydrogen in the CRL forming region. Based on the examination of
230: the models for W48A and our experience in modeling other CRL
231: data set (\nocite{retal05a}Roshi \etal\ 2005;
232: \nocite{retal06}Roshi \etal\ 2006),
233: it is expected that the values for temperature in Table~\ref{tab1}
234: can vary by a factor of 2 and the values for
235: neutral density and $L_{\parallel}$
236: can vary by a factor of 4. The model parameters are derived using
237: a simplified geometry of plane parallel PDR slabs.
238: A comparison of the PDR properties
239: obtained from such models toward W48A with those obtained from models which
240: take into account the photo and chemical processes in the PDR
241: shows that detailed modeling gives values within the factors quoted above
242: (Jeyakumar \etal\ 2007 in preparation).
243:
244: \section{\alfven speed in molecular clouds and non-thermal width
245: of carbon recombination lines}
246: \label{sec3}
247:
248: Molecular clouds are largely neutral with a small admixture of
249: partially ionized heavy elements with their free electrons
250: (ionization fraction 10$^{-5}$). Any wave motion of the magnetic fields
251: in molecular clouds are strongly coupled to the ions.
252: For those waves with $\lambda$ \gsim~$\lambda_{min}$, the time scale
253: of momentum transfer between ions and neutrals is smaller than
254: the magnetic perturbation time scale and hence neutrals
255: are also coupled to such waves (\nocite{am75}Arons \& Max 1975). For typical
256: magnetic fields of 1 mG near `dense' regions in star forming clouds
257: and neutral densities of 10$^6$ \cmthree\
258: (eg. \nocite{jmn89}Johnston, Migenes \& Norris 1989),
259: $\lambda_{min} \sim 3 \times 10^{-7}$ pc, which is
260: much smaller than the size of these dense regions (fraction
261: of a parsec). Thus magnetic waves with $\lambda$
262: \gsim~$3 \times 10^{-7}$ pc can exist in dense regions
263: and the characteristic speed (\alfven speed) of these
264: waves is determined by the total density (ie ion + neutral density)
265: of the molecular cloud (\nocite{am75}Arons \& Max 1975).
266:
267: \nocite{c99}Crutcher (1999) compiled the available sensitive Zeeman measurements
268: of magnetic field strengths in molecular clouds and their neutral densities. Using
269: this data, I
270: estimate the \alfven speed $V_A$ (units of cm s$^{-1}$)
271: in these clouds;
272: \be
273: V_A = \frac{2 B_{los}}{\sqrt{4\pi\mu n_H m_H}},
274: \label{eq:valph}
275: \ee
276: where $n_H$ is the hydrogen atom density in units of \cmthree,
277: $\mu = 1.4$ is the effective mass of an H+He gas with cosmic abundance
278: and $m_H$ is the mass of the hydrogen atom in gm. The measured
279: line-of-sight magnetic field, $B_{los}$, in units of G,
280: is multiplied by 2 to convert
281: it into total field strength (\nocite{c99}Crutcher 1999). Fig.~\ref{fig2}
282: shows the estimated \alfven speed in the molecular cloud sample
283: taken from Crutcher (1999). The estimated \alfven speeds are
284: confined to a narrow range
285: between 0.7 and 4 \kms\ with a median value of 1.6 \kms. This ``constancy'' of
286: \alfven speed was noted earlier and has been understood from models
287: of magnetic confinement of molecular clouds
288: (eg. \nocite{mp95}Moucschovias \& Psaltis 1995; \nocite{b2000}Basu 2000).
289:
290:
291: We now compare the non-thermal velocity dispersion of carbon lines,
292: $\sigma_{NT}$, observed from
293: PDRs with \alfven speed in molecular clouds. The non-thermal width
294: is obtained by removing the
295: thermal contribution, estimated using the gas temperature
296: $T_{PDR}$ (see Table~\ref{tab1}), from the observed CRL width.
297: The non-thermal line widths are plotted in Fig.~\ref{fig2} for the corresponding PDR
298: densities inferred from carbon line modeling.
299: The non-thermal widths have values
300: between 1.6 and 6 \kms with a median value of 2.9 \kms.
301: Fig.~\ref{fig2} shows
302: that the non-thermal line widths are almost ``constant'' over three orders
303: of magnitude in density. The ``constancy'' of the non-thermal
304: line widths and their magnitudes are similar (less than a factor of 2)
305: to those inferred for \alfven waves in molecular clouds.
306:
307:
308: The gas phase carbon is ionized in PDR and hence is
309: strongly coupled to the magnetic field in these regions.
310: If we consider similar physical parameters in PDRs as in
311: dense molecular regions ($n_H \sim 10^6$ \cmthree; $B \sim$ 1 mG)
312: then $\lambda_{min}$ of $\sim 3 \times 10^{-7}$ pc is
313: at least two orders of magnitude smaller than the typical
314: line-of-sight thickness of the PDR (ie $L_{\parallel}$).
315: Thus \alfven waves with $\lambda$ \gsim~$\lambda_{min}$
316: exist in the PDR and here I consider their contribution to
317: the non-thermal width of the observed carbon lines. The amplitude of
318: the velocity of carbon ions, $\delta v$, due to these waves
319: is related to the magnetic perturbation amplitude
320: $\delta B$ through the equation (eg. \nocite{am75}Arons \& Max 1975)
321: \be
322: \delta v = \frac{\delta B}{\sqrt{4\pi\mu n_H m_H}}.
323: \ee
324: The observed non-thermal velocity dispersion is approximately
325: given by $\delta v$. It is usually assumed
326: that $\delta B \sim B$, in that case, the right hand side of
327: Eq. (2) becomes identical to that of Eq.~(\ref{eq:valph}).
328: Based on these considerations and the
329: characteristics exhibited by the non-thermal velocity dispersion
330: and \alfven speed in molecular clouds (see above) we assume that the
331: velocity dispersion of carbon lines is
332: an estimate of the \alfven speed in the PDR.
333:
334: The observed velocity dispersion is usually scaled by $\sqrt{3}$
335: to convert it to 3D velocity dispersion. This scaling assumes random magnetic
336: field orientation along line-of-sight and random polarization
337: of \alfven waves. Carbon lines are observed from PDRs near \UCHII s
338: where shocks are present. In such shocked regions only the tangential
339: component of the magnetic field is amplified and the
340: \alfven speed associated with this component is scaled
341: by the square root of the density compression ratio
342: (\nocite{mz95}McKee \& Zweibel 1995). Hence the
343: scaling factor needed to convert the observed velocity dispersion
344: to a 3D dispersion is uncertain. Direct observation
345: of the Zeeman effect of CRLs may help in determining
346: this factor (see Section~\ref{sec5}). Here I note a systematically
347: high value (a factor of 1.8) for the CRL velocity dispersion
348: compared to the \alfven speed in molecular clouds,
349: which may be an indication of higher \alfven speeds
350: in PDR shocks.
351:
352: \section{Magnetic field in Photo-dissociation regions}
353: \label{sec4}
354:
355: The magnetic field strength is obtained using Eq~(\ref{eq:valph})
356: by substituting the estimated non-thermal velocity dispersion of CRL for
357: the \alfven speed and using the neutral density obtained
358: from modeling the CRL emission
359: (eg. \nocite{retal05b}Roshi \etal\ 2005b). Field strength values
360: thus obtained are tabulated in Table.~\ref{tab1}. These values
361: represent the total magnetic field strength in the PDR.
362: Based on the expected range of the derived physical properties
363: of the PDR (see Section~\ref{sec2}), the estimated uncertainty
364: in $B$ is typically a factor of 2.5.
365:
366:
367: In Fig.~\ref{fig3}, we compare the estimated magnetic field with
368: those measured toward molecular clouds. The ordinate of the plot
369: is the number density of $H_2$ molecules. Here, the estimated
370: field strength values in the PDR are compared with those measured
371: in molecular
372: clouds with similar density. Such a comparison is possible since
373: earlier observations toward molecular clouds show that the magnetic
374: field scales with density ($n_{H_2} \propto \rho^{0.47}$;
375: \nocite{c99}Crutcher 1999). To produce Fig.~\ref{fig3},
376: the neutral density of the PDR given in Table.~\ref{tab1} is divided by 2
377: to convert it into number density of H$_2$. As seen in the figure
378: the estimated magnetic field strengths in the PDR compare well within errors with
379: those measured toward molecular clouds with similar density.
380:
381: Magnetic field measurements using Zeeman effect of CRLs near 1.4 GHz
382: were attempted toward a few \HII\ regions (\nocite{s84}Silverglate 1984)
383: of which W48 and S88B are of interest here. The 1.4 GHz CRL
384: emission toward W48 does not originate from the PDR associated
385: with the \UCHII\ (\nocite{retal05a}Roshi \etal\ 2005a) and
386: hence a comparison of the upper limit on the field strength obtained by
387: \nocite{s84}Silverglate (1984) with the estimated value here is
388: not meaningful. Toward S88B, the upper limit for the magnetic
389: field strength obtained is consistent with estimated
390: values given in Table~\ref{tab1}. Magnetic field
391: measurements using OH or \HI\ Zeeman effects are available
392: in literature toward a few \UCHII s listed in Table~\ref{tab1}.
393: I compare the field strength given in Table~\ref{tab1}
394: with the results of these observations. Note that OH and \HI\
395: lines may originate from different
396: spatial locations compared to the CRL forming region.
397: From OH Zeeman observations toward S88B a field strength
398: in the range 0.1 -- 0.3 mG was obtained by Sarma \etal\ (2006 in preparation)
399: consistent with our estimate.
400: \nocite{vg90}Van der Werf \& Goss (1990) using \HI\ Zeeman observations
401: measured a peak magnetic field of 0.1 mG in the $-$45 \kms
402: component observed toward W3. The CRL LSR velocity is comparable with this velocity
403: and the field strength is consistent with the estimated value.
404: \nocite{bt01}Brogan \& Troland (2001) measured a maximum
405: field strength of 0.3 mG toward W49A. They used Zeeman effect
406: of \HI\ to measure the field strength with an angular resolution
407: of 25\arcsec. The measured value is about twenty times smaller than
408: the estimated value; possibly because the two tracers (\HI\ and CRL)
409: do not probe the same region in this case.
410:
411: \section{Discussion}
412: \label{sec5}
413:
414: As mentioned in Section~\ref{sec2}, CRL emission is dominated by stimulated
415: emission for the transitions listed in Table~\ref{tab1}.
416: Because of the dominance of stimulated emission, carbon line
417: is detected only from the near side of the \UCHII. The non-detection
418: from far side of \UCHII\ means that the width of CRL is not contributed
419: to by the expansion of the \UCHII\ (if the \UCHII\ were expanding).
420: Thus the line width has contribution only from thermal and non-thermal motions.
421: The carbon line width may also have contributions from large scale motions such
422: as outflows. However, the similarity of line widths observed from different sources
423: (see Table~\ref{tab1}) may indicate that this contribution is small.
424: Thus the field strength estimated in Section~\ref{sec4} may not
425: be affected by such large scale motions. However, presence of
426: any non-magnetic turbulence would affect the estimation of the field strength and
427: hence the values obtained should be considered as upper limits.
428:
429: %To estimate the magnetic field, we used the non-thermal line width
430: %without multiplying by $\sqrt{3}$. This multiplicative factor is used
431: %in the literature to convert the measured one dimensional Doppler
432: %velocity to the magnitude of velocity in three dimension assuming
433: %that the velocity distribution is random. I haven't used this scaling
434: %%factor since the line-of-sight extent of the CRL forming region
435: %is of the order of 10$^{-3}$ pc (see Table~\ref{tab1}) and the
436: %observations may approximately be measuring the total velocity of the
437: %carbon ion motion in the magnetic field. However, this assumption
438: %needs further observational verification.
439:
440: The magnetic origin of CRL width can be confirmed by
441: measuring the field strength using Zeeman effect of carbon lines
442: from the PDRs and comparing it with the
443: values estimated in Section~\ref{sec4}. If confirmed then
444: multi-frequency CRL observations form another tool to deduce
445: the strength of magnetic fields near star forming regions.
446:
447:
448: \acknowledgments
449:
450: I acknowledge many useful discussions with K. Subramanian, S. Sridhar,
451: A. A. Deshpande, D. Bhattacharya and helpful comments from an anonymous referee.
452:
453: \begin{thebibliography}{}
454: \bibitem{am75}Arons, J., Max, C. E., 1975, ApJ, L77
455: \bibitem{bmw64}Barrett, A. H., Meeks, M. L., Weinreb, S., 1964, Nature, 202, 475
456: \bibitem{b2000}Basu, S., 2000, ApJ, 540, L106
457: \bibitem{bt01}Brogan, C. L., Troland, T. H., 2001, ApJ, 550, 799
458: \bibitem{c02}Churchwell, E., 2002, ARA\&A, 40, 27
459: \bibitem{cetal90}
460: Churchwell, E., Walmsley, C. M., Cesaroni, R., 1990, A\&AS, 83, 119
461: \bibitem{c99}Crutcher, R. M., 1999, ApJ, 520, 706
462: %\bibitem{e99}Evans, N. J. II, 1999, ARA\&A, 37, 311
463: \bibitem{getal98}Garay, G., Lizano, S., Gomez, Y., Brown, R. L., 1998,
464: ApJ, 501, 699
465: \bibitem{goetal98}Gomez, Y., \etal\ , 1998, ApJ, 503, 297
466: %\bibitem{hetal93}Heiles, C., \etal\, 1993, Protostars and planets III, Eds E. H. Levy
467: %\& J. I. Lunine, University of Arizona Press, Tucson, p279
468: \bibitem{ht97}
469: Hollenbach, D. J., Tielens, A. G. G. M., 1997, ARA\&A, 35, 179
470: \bibitem{jmn89}Johnston, K. J., Migenes, V., Norris, R. P., 1989, ApJ, 341, 847
471: \bibitem{kag98}Kantharia, N. G., Anantharamaiah, K. R., Goss, W. M., 1998, ApJ, 504, 375
472: \bibitem{l81}Larson, R. B., 1981, MNRAS, 194, 809
473: %\bibitem{letal89} Leisawitz, D., Bash, F. N., Thaddeus, P., 1989, ApJS, 70, 731
474: \bibitem{mz95}McKee, C. F., Zweibel, E. G., 1995, ApJ, 440, 686
475: \bibitem{m03}MacLow, M. M., in Simulations of magnetohydrodynamic turbulence
476: in astrophysics, eds. T. Passot \& E. Falgarone, Lecture Notes in Physics, 614, 182
477: \bibitem{metal74}Morris, M., Zuckerman, B., Turner, B. E., Palmer, P., 1974, ApJ,192, L27
478: \bibitem{m75}Mouschovias, T. Ch., 1975, Ph.D. thesis, Univ. California, Berkeley
479: %\bibitem{mm85} Mouschovias, T. Ch., Morton, S. A., 1985, ApJ, 298, 190
480: \bibitem{m87} Mouschovias, T. Ch., 1987, Physical processes in interstellar clouds,
481: Eds G. E. Morfill \& M. Scholer, D. Reidel publishing company, Dordrecht, p453
482: \bibitem{mp95}Mouschovias, T. Ch., Psaltis, D., 1995, ApJ, 444, L105
483: %\bibitem{m83}Myers, P. C., 1983, ApJ, 270, 105
484: \bibitem{mg88a}Myers, P. C., Goodman, A. A., 1988a, ApJ, 326, L27.
485: \bibitem{mg88b}Myers, P. C., Goodman, A. A., 1988b, ApJ, 329, 392.
486: \bibitem{nwt94}Natta, A., Walmsley, C. M., Tielens, A. G. G. M., 1994, ApJ, 428, 209
487: \bibitem{retal05a}Roshi, D. A., Goss, M. W., Anantharamaiah, K. R., Jeyakumar, S., 2005a, ApJ, 626, 253
488: \bibitem{retal05b}Roshi, D. A., \etal\ 2005b, ApJ, 625, 181
489: \bibitem{retal06}Roshi, D. A., De Pree, C. G., Goss, W. M, Anantharamaiah, K. R., 2006, ApJ, 644, 279
490: \bibitem{sal87}Shu, F. H., Adams, F. C., Lizano, S., 1987, ARA\&A, 25, 23
491: \bibitem{s84}Silverglate, P. R., 1984, ApJ, 279, 694
492: %\bibitem{st78} Silverglate, P. R., Terzian, Y., 1978, ApJ, 224, 437
493: %\bibitem{s78}Spitzer, L. Jr., 1978, Physical properties in the Interstellar Medium (New York: Wiley-Interscience)
494: \bibitem{th86}Troland, T. H., Heiles, C., 1986, ApJ, 301, 339
495: \bibitem{v89}Vallee, J. P., 1989, A\&A, 224, 191
496: \bibitem{vg90}van der Werf, P. P., Goss, W. M., 1990, A\&A, 238, 296
497: \bibitem{ze74}Zuckerman, B., Evans, N. J. II, 1974, ApJ, 192, L149
498: %\bibitem{zp74}Zuckerman, B., Palmer, P., 1974, ARA\&A, 12, 279
499: \end{thebibliography}
500:
501: \clearpage
502:
503: \begin{figure}
504: \includegraphics[width=3.0in, height=3.3in, angle = -90]{fig1.ps}
505: \caption{Examples of carbon recombination line spectra toward
506: \UCHII s G43.24$-$0.05 and G70.33$+$1.59 observed near 9 GHz
507: (Roshi \etal\ 2005a). The Gaussian profiles plotted
508: in dotted line correspond to an assumed PDR gas temperature of 1000 K.
509: The large difference between the observed line profiles and the Gaussian
510: curves demonstrates the dominance of non-thermal motions in the
511: PDR. The LSR velocity is with respect to the C89$\alpha$
512: (9.1779 GHz) transition.}
513: \label{fig1}
514: \end{figure}
515:
516: \clearpage
517:
518: %anish
519:
520: \begin{figure}
521: \includegraphics[width=3.0in, height=3.3in, angle = -90]{fig2.ps}
522: \caption{\alfven speed in molecular clouds (crosses) and non-thermal
523: velocity dispersion of carbon lines (triangles) observed toward PDRs
524: near \UCHII s are plotted against density in these regions.
525: \alfven speeds are estimated from the molecular line data compiled
526: by Crutcher (1999). Errors in the non-thermal velocity dispersion
527: are derived from the uncertainties in the
528: measurements and estimation of the model parameters.
529: The expected range of the estimated densities for the PDR is shown by
530: the horizontal bar. The plot shows that
531: the non-thermal component of the CRL width is ``constant''
532: for almost all PDRs and has a median value similar to that
533: of \alfven speed in molecular cloud.}
534: \label{fig2}
535: \end{figure}
536:
537: \clearpage
538:
539: \begin{figure}
540: \includegraphics[width=2.5in, height=3.3in, angle = -90]{fig3.ps}
541: \caption{Log-log plot of the magnetic field strength vs neutral density.
542: The plot compares the magnetic fields measured in molecular clouds
543: (crosses; data taken from Crutcher 1999) with those estimated
544: using recombination line data (triangles). Comparison is
545: done for similar $H_2$ densities in the
546: molecular cloud and the PDR. Such a comparison is possible since
547: earlier observations toward molecular clouds show that the magnetic
548: field scales with density (Crutcher 1999).
549: The plot shows that the field strengths
550: estimated using CRL data (see Section.~\ref{sec4})
551: are comparable with those
552: measured in molecular clouds with densities similar to
553: PDR densities.
554: The expected range of density
555: and estimated magnetic field are shown by horizontal and
556: vertical error bars respectively. }
557: \label{fig3}
558: \end{figure}
559:
560: \clearpage
561:
562: \input{tab1.tex}
563:
564: \end{document}
565: