1: \documentclass[12pt,preprint]{aastex}
2:
3: %\usepackage{emulateapj5}
4:
5: %\usepackage{psfig}
6:
7:
8: \def\kms{\relax \ifmmode {\,\rm km~s}^{-1}\else \,km~s$^{-1}$\fi}
9: \def\cm-3{\relax \ifmmode {\,\rm cm}^{-3}\else \,cm$^{-3}$\fi}
10: \def\deg{\hbox{$^\circ$}}
11: \def\arcsec{\hbox{$^{\prime\prime}$}}
12: %\def\n2h+ {N$_2$H$^+$}
13: \def\h13cn{H$^{13}$CN}
14: \def\h13co+{H$^{13}$CO$^+$}
15:
16:
17: \begin{document}
18:
19: \title{
20: High Angular Resolution Measurement of Ion and Neutral Spectra as a Probe
21: of the Magnetic Field Structure in DR21(OH)}
22:
23: \author{Shih-Ping Lai\altaffilmark{1}, T. Velusamy, and W. D. Langer\\
24: {\it slai@astro.umd.edu, velu@jpl.nasa.gov, william.langer@jpl.nasa.gov}}
25: \affil{Jet Propulsion Laboratory, California Institute of Technology, MS
26: 169-506, 4800 Oak Grove Dr., Pasadena, CA 91109}
27: \altaffiltext{1}{Current address: Department of Astronomy, University of
28: Maryland, College Park, MD 20742-2421}
29:
30: \begin{abstract}
31: It has been suggested that under average interstellar field
32: strengths the cyclotron interaction between ions and magnetic fields
33: is strong enough to narrow the linewidth and suppress the line wings
34: in the ion spectra. We present evidence for the
35: cyclotron interaction effect at arcsec scale on the
36: velocity dispersions in the spectra of ion/neutral molecular species
37: in DR21(OH) observed with the OVRO-MMA.
38: Using a spatial resolution $\sim$ 3 times higher than previous
39: CSO observations by Houde et al.\ (2002),
40: we show that H$^{13}$CO$^+$ and H$^{13}$CN are coexistent at the scale
41: of our observations (6\arcsec).
42: In the eastern parts of the DR21(OH) core where the dynamics is simple,
43: the ion linewidths are indeed narrower than the neutral linewidths with an
44: average ion-to-neutral linewidth ratio of 0.82$\pm$0.04.
45: We use our results, along with the existing Zeeman and dust/CO polarization
46: data on small scales, to derive the 3-D magnetic field structure.
47: We obtain a field strength of $0.44\pm$0.12 mG with inclination of 36\deg\
48: to the line of sight, directed toward the observer, and a position angle of
49: $-$75\deg\ in the plane of the sky.
50: With the full magnetic field strength derived here,
51: we are able to conclude that the MM1 core of DR21(OH) is magnetically
52: supercritical; although turbulence provides the dominant support.
53:
54: \end{abstract}
55: \keywords{star formation, magnetic fields}
56:
57:
58: \section{Introduction}
59:
60: Measurements of magnetic fields in the environment of star forming regions
61: are mostly carried out with observations
62: of Zeeman splitting (Crutcher 1999),
63: the polarimetry of dust continuum (Dotson et al.\ 2000;
64: Lai et al.\ 2001, 2002), and the polarimetry of CO emission
65: (Girart et al.\ 1999; Lai et al.\ 2003).
66: The Zeeman observations only give the field {\it strength} along the line of
67: sight, and the polarimetry of dust continuum and CO emission
68: yield the field {\it direction} in the plane of sky.
69: As a result, these observations do not determine
70: the full strength and orientation of the magnetic fields in 3-D.
71: To characterize the magnetic field environment,
72: we need to measure directly the magnetic field strength in the plane of sky
73: $B_p$.
74: Without direct measurement of the full field strengths,
75: the role magnetic fields play in star formation can
76: only be assessed using indirect estimates of $B_p$
77: (the Chandrasekhar-Fermi method as used by Lai et al.\ 2001)
78: and some statistical arguments (Crutcher 1999).
79:
80: Houde et al.\ (2002) proposed an interesting
81: new approach to probe the orientation of magnetic fields.
82: Houde et al.\ (2000a,b) show that in weakly ionized plasma,
83: even under weak fields ($\sim 10~\mu$G),
84: the cyclotron interaction between the ions and the field can
85: reduce the linewidth and suppress the high velocity wings in the ion
86: spectra when compared to coexistent neutral spectra.
87: Their HCN and HCO$^+$ observations with the Caltech Submillimeter
88: Observatory (CSO) show evidence of these predicted features.
89: Because the ion-to-neutral linewidth ratio depends largely on the
90: inclination angle of the magnetic fields (no cyclotron interaction
91: along the field lines),
92: the full strength and orientation of magnetic fields can
93: be obtained by combining the observations of Zeeman splitting,
94: polarimetry of dust continuum and/or CO emission,
95: and the ion-to-neutral linewidth ratio.
96: Houde et al.\ (2002) have constructed the 3-D field structure of M17
97: using this method.
98:
99: However, it is difficult to interpret
100: the differences in the velocity dispersions of the ion/neutral
101: spectra unambiguously as being caused by the cyclotron interaction,
102: if the beam is large enough to include the spectral line
103: emissions from physically and chemically different components.
104: The intrinsic difference between the velocity structure of
105: the ions and neutrals will also contribute to
106: the ion-to-neutral linewidth ratio.
107: For example, if a velocity gradient exists in a large region with HCN,
108: but HCO$^+$ only coexists with HCN in a small part of this region,
109: the HCN linewidth is naturally larger than the HCO$^+$ linewidth.
110: Therefore, it is important to obtain high spatial resolution
111: spectral line maps to resolve the regions with widely
112: different physical and chemical characteristics.
113:
114: In this paper, we present the comparison between
115: H$^{13}$CN and H$^{13}$CO$^+$ spectra in DR21(OH) observed with
116: the millimeter array of the Owens Valley Radio Observatory (OVRO).
117: DR21(OH) is an excellent object to study the magnetic
118: field structure, because it is in the early stage of massive star formation.
119: DR21(OH) is one of the few sources with measured average line-of-sight
120: magnetic field strength from CN Zeeman detections (Crutcher et al.\ 1999)
121: and field directions in the plane of sky mapped with BIMA
122: using both dust and CO polarimetry (Lai et al.\ 2003).
123: Deriving the full strength and 3-D structure requires knowledge of the
124: inclination angle of the magnetic field to
125: the line of sight, which can be obtained by measuring the
126: ion-to-neutral linewidth ratio.
127:
128: \section{Observations and Results}
129:
130: The observations were made with the OVRO Millimeter Array
131: at 3mm in the Compact, Low, and High resolution array configurations
132: between October 2002 and May 2003. We obtained simultaneously spectral line
133: images of HCN \& HCO$^+$ and H$^{13}$CN \& H$^{13}$CO$^+$;
134: $^{13}$CS and N$_2$H$^+$ were observed in the opposite sidebands.
135: The synthesized beam is 6.4\arcsec$\times$6.9\arcsec\ for natural weighting
136: and spectral resolutions are 0.434 \kms.
137: Among these images, the HCN and HCO$^+$ spectra show features of
138: optically thick emission with strong self-absorption;
139: N$_2$H$^+$ has a very different spatial distribution
140: compared to other species; H$^{13}$CN, H$^{13}$CO$^+$, and $^{13}$CS
141: have a similar spatial distribution, but $^{13}$CS has a different
142: velocity structure from the other two. Therefore,
143: H$^{13}$CN and H$^{13}$CO$^+$ are the only neutral/ion pair that
144: can be used to examine the cyclotron effect in our data.
145:
146: Figure 1 shows the 3mm continuum and the integrated flux of
147: H$^{13}$CN and H$^{13}$CO$^+$. The two compact cores in DR21(OH),
148: MM1 and MM2, are resolved. H$^{13}$CN and H$^{13}$CO$^+$ both
149: trace the dense cores very well. H$^{13}$CO$^+$ is slightly more
150: extended than H$^{13}$CN, which is probably due to the lower
151: critical density of H$^{13}$CO$^+$. Therefore,
152: H$^{13}$CN and H$^{13}$CO$^+$ seem to occupy the same volume.
153: Figure 2 shows the normalized H$^{13}$CN and H$^{13}$CO$^+$
154: spectra overlaid on their flux ratio map.
155: Note that H$^{13}$CN has three hyperfine lines with intensity ratios
156: of 1:5:3 for the optically thin case. The spectra show that DR21(OH)
157: contains two dominant velocity components -- one at $\sim -5$\kms\ and
158: the other at $\sim -1$\kms, which are likely to be
159: associated with MM1 and MM2, respectively.
160: At the positions between MM1 and MM2, the spectra are more complex
161: and may contain
162: an absorption feature at $\sim -$2.5\kms. Overall, the spectra
163: of H$^{13}$CN and H$^{13}$CO$^+$ trace each other well, which is strong
164: evidence for the coexistence of H$^{13}$CN and H$^{13}$CO$^+$,
165: as also suggested by their spatial intensity correlation.
166:
167: A difference in the ion/neutral linewidth can arise in some cases from
168: chemical
169: inhomogeniety, especially, if multiple velocity features are present. In Figure
170: 2 we show the distribution of the H$^{13}$CN-to-H$^{13}$CO$^+$
171: flux ratio (in greyscale) which is a measure of the chemical
172: differentiation in the core. We can safely
173: discard the effects of chemical differentiation by excluding regions where the
174: flux ratio changes
175: appreciably over the small spatial scale sizes of the OVRO beam. However,
176: chemical
177: differentiation alone will not always cause linewidth differences. The velocity
178: structure of the region also needs to be considered.
179: For example, if the velocity structure is uniform across the cores
180: such that the shape and the central velocity of the spectra stay the same,
181: any local variation in the ion-to-neutral flux ratio will result in a weighting
182: factor that affects only the line intensities, not the
183: ion-to-neutral linewidth ratios
184: for the whole core.
185: However, in DR21(OH) at least two velocity components are detected and
186: therefore,
187: it is critical to isolate the regions with different dynamics when
188: interpreting the ion-to-neutral linewidth ratio as the cyclotron effect.
189: We see a large variation in their flux ratio (by factors of 0.5 to 4) over the
190: entire region,
191: but it appears to be smooth except near the southwest parts of the core.
192: In the eastern parts of the
193: core the variation in the flux ratio is small,
194: and the fully resolved velocity structure is dominated by a single feature.
195: Thus, in this region of DR21(OH), we are able to
196: minimize any contribution to the ion/neutral linewidth differences arising
197: from chemical inhomogeneity. In our analysis we do not include the southwest
198: region where rapid and significant variation of ion-to-neutral flux ratio is
199: observed.
200: Furthermore, the ion/neutral spectra in the southwest region have low S/N.
201:
202:
203: In the eastern part of MM1 where the $-$5\kms\ component dominates and
204: the spectral line shapes are better defined,
205: we can see that H$^{13}$CO$^+$ is slightly narrower
206: than H$^{13}$CN. To eliminate the complication of the H$^{13}$CN
207: hyperfine lines, we simulate H$^{13}$CN spectra with all three hyperfine
208: lines using the H$^{13}$CO$^+$ as a `template'.
209: For the case of a single velocity component,
210: if the linewidth of H$^{13}$CN is smaller than
211: half of the velocity separations between hyperfine lines,
212: the difference spectrum between the observed H$^{13}$CN and
213: that simulated from H$^{13}$CO$^+$ is
214: largely free from the confusion caused by the hyperfine components
215: and thus is a useful measure of the cyclotron interaction.
216: If indeed H$^{13}$CN has larger linewidths than H$^{13}$CO$^+$,
217: the difference spectra will have two positive peaks near
218: the wings of the line profile as demonstrated in
219: Figure 2d. As the line shapes of H$^{13}$CN and H$^{13}$CO$^+$
220: are approximately Gaussians, the integration of the difference spectra over
221: the extent of the main hyperfine component gives a measure of
222: the difference of the dispersions in the spectra.
223: \begin{equation}
224: \int{(I_n - I_i)}{dv} = \sqrt{2\pi} (\sigma_n - \sigma_i),
225: \end{equation}
226: \noindent where $I_n$ and $I_i$ are the normalized H$^{13}$CN and
227: H$^{13}$CO$^+$ spectra and $\sigma_n$ and $\sigma_i$
228: are the dispersions of H$^{13}$CN and H$^{13}$CO$^+$, respectively.
229: Figure 3a shows the map of $(\sigma_n - \sigma_i)$ which is calculated by
230: integrating the difference spectra
231: over the $-5$\kms\ component ($v_{LSR}=-9$ to $-2.5$\kms).
232: These $(\sigma_n - \sigma_i)$ estimated using Eq(1) are in good
233: agreement (within 10\%) with those obtained by fitting Gaussians to
234: the observed H$^{13}$CN and H$^{13}$CO$^+$ line profiles at (+10\arcsec,
235: +5\arcsec).
236: The signal-to-noise ratio (S/N) of $(\sigma_n - \sigma_i)$
237: is presented in Figure 3b.
238: We use $(\sigma_n - \sigma_i)$ and $\sigma_i$
239: measured from H$^{13}$CO$^+$
240: spectra to derive $(\sigma_i/\sigma_n)$ in Figure 3c.
241: It is evident from the results shown in Figure 3 that in the eastern
242: side of MM1 the ion linewidth is smaller than the neutral linewidth.
243: We do not present similar analysis for the $-1$\kms\ component,
244: because of the low S/N of $(\sigma_n - \sigma_i)$.
245:
246:
247: \section{Discussion}
248:
249: The interferometer observations with OVRO have the advantage to resolve
250: the velocity and chemical structures in the DR21(OH) core with high spatial
251: resolution. In the eastern parts of DR21(OH)
252: our observations show that H$^{13}$CN and H$^{13}$CO$^+$ coexist over ~15\arcsec
253: $\times$10\arcsec\ in a single velocity feature
254: and confirm that
255: the ion line is indeed narrower than the neutral line.
256: We use the ion-to-neutral linewidth ratio $(\sigma_i/\sigma_n)$ shown in Figure
257: 3c
258: to infer the inclination angle of the magnetic field to
259: the line of sight ($\alpha$) according to Equation (11)
260: of Houde et al.\ (2002). To derive the value of ($\alpha$) using the
261: ion/neutral linwidth ratio we require knowledge of the gas motions within
262: the core.
263: Houde et al.\ consider the cases with neutral flow
264: collimation widths, $\Delta\theta$, ($\Delta\theta=90$\deg\ means no
265: collimation as in a purely turbulent case). For a given value of
266: $(\sigma_i/\sigma_n)$ a range of values for $\alpha$ are possible depending
267: on the values for flow collimation widths (see Figure 4 in Houde et al.). For
268: example the
269: average ion-to-neutral linewidth ratio, $(\sigma_i/\sigma_n)$ of 0.82 will
270: correspond to an inclination angle $\alpha$ = 76\deg, 60\deg and 36\deg\
271: for $\Delta\theta$=20\deg, 45\deg and 90\deg\ respectively.
272: The dependence of the estimates for $\alpha$ on $\Delta\theta$ is less critical
273: for $\Delta\theta \sim$ 90\deg, a case when turbluent motions dominate over
274: collimated flows.
275: Although MM1 in DR21(OH) does power high velocity CO outflows (Lai et al.\
276: 2003),
277: there is no evidence that the spectra of ions and neutrals observed
278: at the rest velocity of the cores are contaminated by collimated flows.
279: Therefore, to derive the magnetic field inclination, we have
280: ignored collimated flows and we assume the gas is fully
281: turbulent ($\Delta\theta=90$\deg).
282: Figure 4 shows the inclination angle $\alpha$ ranges from 30\deg\
283: in the northern part of MM1 to 42\deg\ in the south, and the uncertainty in
284: $\alpha$ is $\sim$ 4--8\deg. The variation in $\alpha$ is small (12\deg) and
285: is comparable to the variation of
286: the overlaid $B_p$ vectors derived from the dust polarization
287: (black) and CO polarization (white).
288: Therefore, the magnetic field directions seem to be approximately
289: uniform across the MM1 region.
290: The average line-of-sight magnetic field strength in
291: MM1 $\sim$ $-$0.36$\pm$0.10 mG as measured by
292: CN Zeeman observations (Crutcher et al.\ 1999).
293: With the average $\alpha=36$\deg,
294: the field strength in the plane of sky ($B_p$) is 0.25$\pm$0.07 mG, and
295: the full magnetic field strength is $ 0.44\pm$0.12 mG. The direction of the
296: field is out of the plane of the sky towards the observer at an angle of
297: 36\deg to the line of sight, and its projection on the sky is at a
298: position angle $-75$\deg.
299:
300: Lai et al.\ (2003) estimate $B_p\sim$0.9 mG for MM1 from the observed
301: dispersion in the CO polarization position angles using
302: the Chandrasekhar-Fermi formula.
303: This value is a factor of 3 larger than what is derived from the cyclotron
304: interaction data presented here. This discrepancy
305: could be partly due to the low values for the dispersion caused by the
306: smoothing of polarization by the finite beam size of the BIMA observations.
307: Also, the `calibration factor' of the Chandrasekhar-Fermi formula (0.5)
308: used by Lai et al.\ (2003) is an average value of clouds
309: in numerical simulations (Ostriker, Gammie, \& Stone 2001),
310: which may not be a proper value DR21(OH).
311:
312: With the knowledge of the full magnetic field strength, we can evaluate
313: the role that magnetic fields play for star formation in MM1.
314: First, we can calculate the mass-to-magnetic flux ratio
315: from $M/\Phi_B=1.0\times 10^{-20}N(H_2)/|B|$
316: in units of the critical mass-to-flux ratio (Crutcher 1999),
317: where $N(H_2)$ is the column density in cm$^{-2}$
318: and $|B|$ is the field strength in $\mu$G.
319: With $N(H_2)=2\times 10^{23}$ cm$^{-2}$
320: and B=0.44 mG, $M/\Phi_B=$4.5.
321: Therefore, MM1 is clearly magnetically supercritical.
322: However, turbulence seems to provide dominant support.
323: The H$^{13}$CO$^+$ linewidth ($\sim$ 2.2\kms) is much
324: larger than its thermal linewidth for $T_k$=50 K (0.3\kms),
325: indicating strong turbulence in DR21(OH).
326: The relative importance of magnetic, thermal, and turbulent support
327: can be evaluated by comparing their energy to the gravitational
328: energy.
329: We find that in MM1 the magnetic-to-gravitational energy ratio
330: is 0.05, the thermal-to-gravitational energy ratio is 0.16,
331: and the turbulent-to-gravitational energy ratio is 0.66.
332: Therefore, turbulence provides the dominant support in MM1.
333: Although our conclusion on the status of MM1 is similar to
334: what has been suggested by Crutcher et al.\ (1999)
335: with statistical argument for the full magnetic field strength,
336: our results exclude the possiblity of the magnetic field dominance
337: which may be true only if the field orientation in 3-D is very close to
338: the plane of the sky.
339:
340: \acknowledgements
341: We thank the referee for helpful suggestions. This research was conducted at the
342: Jet Propulsion Laboratory, California Institute of Technology, with support from
343: the National Aeronautics and Space Administration,
344: while SPL held a National Research Council Resident Research Associateship
345: at JPL. The OVRO Millimeter Array is supported by NSF grant AST-99-81546.
346:
347: \begin{thebibliography}{}
348: \bibitem[]{302} Crutcher, R.~M.\ 1999, ApJ, 520, 706
349: \bibitem[]{303} Crutcher, R.~M., Troland, T.~H., Lazareff, B., Paubert, G., \&
350: Kaz{\`
351: e}s, I.\ 1999, ApJL, 514, L121
352: \bibitem[]{306} Dotson, J.~L., Davidson, J., Dowell, C.~D., Schleuning, D.~A.,
353: \&
354: Hildebrand, R.~H.\ 2000, ApJS, 128, 335
355: \bibitem[]{308} Girart, J.~M., Crutcher, R.~M., \& Rao, R.\ 1999, ApJL, 525,
356: L109
357: \bibitem[]{309} Houde, M., Bastien, P., Peng, R., Phillips, T.~G., \& Yoshida,
358: H.\
359: 2000a, ApJ, 536, 857
360: \bibitem[]{311} Houde, M., Peng, R., Phillips, T.~G., Bastien, P., \& Yoshida,
361: H.\
362: 2000b, ApJ, 537, 245
363: \bibitem[]{313} Houde, M., Phillips, T.~G., Bastien, P., Peng, R., \& Yoshida,
364: H.\
365: 2001, ApJ, 547, 311
366: \bibitem[]{315} Houde, M.~et al.\ 2002, ApJ, 569, 803
367: \bibitem[]{316} Lai, S.-P., Crutcher, R.~M., Girart, J.~M., \& Rao, R.\ 2001,
368: ApJ,
369: 561, 864
370: \bibitem[]{318} Lai, S.-P., Crutcher, R.~M., Girart, J.~M., \& Rao, R.\ 2002,
371: ApJ,
372: 566, 925
373: \bibitem[]{320} Lai, S.-P., Crutcher, R.~M., \& Girart, J.~M.\ 2003, ApJ, in
374: press
375: \bibitem[]{321} Ostriker, E.\ C., Stone, J.\ M., \& Gammie, C.\ F.\ 2001, \apj,
376: 546,
377: 980
378: %\bibitem[Richardson, Sandell, Cunningham, \& Davies(1994)]{1994A&A...286..555R} Richardson, K.~J., Sandell, G., Cunningham, C.~T., \& Davies, S.~R.\ 1994, \aap, 286, 555
379: \end{thebibliography}
380:
381: \begin{figure} % figure 1
382: \plotone{f1.ps}
383: \caption{The 3mm dust continuum and the velocity integrated intensity maps of
384: H$^{13}$CN and H$^{13}$CO$^+$. The lowest contour and the interval are at
385: 10\% and 20\%, respectively, of the peak intensities.}
386: \end{figure}
387:
388: \begin{figure} % figure 2
389: \plotone{f2.ps}
390: \caption{Spectra of Ions and Neutrals.
391: {\bf (a)} Normalized H$^{13}$CN (red) and H$^{13}$CO$^+$ (green)
392: spectra overlaid on H$^{13}$CN-to-H$^{13}$CO$^+$ flux ratio
393: (greyscale). The beam size is indicated. {\bf (b)}
394: H$^{13}$CN (red) and H$^{13}$CO$^+$
395: (green) spectra at offset position (+10$^{\prime\prime}$, +5$^{\prime\prime}$).
396: {\bf (c)} Difference
397: spectrum at this position between observed H$^{13}$CN and simulated H$^{13}$CN
398: (using
399: H$^{13}$CO$^+$ profile as a template - see \S 2) is shown in blue.
400: H$^{13}$CO$^+$ spectrum
401: (green) is shown for comparison. {\bf (d)} Difference
402: spectrum for model Gaussian spectra with $(\sigma_i / \sigma_n)$ = 0.86,
403: equivalent to that for the (+10$^{\prime\prime}$, +5$^{\prime\prime}$)
404: position.}
405: \end{figure}
406:
407: \begin{figure} % figure 3
408: \plotone{f3.ps}
409: \caption[]{Greyscale maps of ion/neutral linewidth differences overlaid on
410: contours of H$^{13}$CO$^+$ integrated flux at 10\%, 30\%, 50\%, 70\%,
411: and 90\% levels. {\bf (a)} Difference of H$^{13}$CN and
412: H$^{13}$CO$^+$
413: line dispersion $(\sigma_n - \sigma_i)$. {\bf (b)} Signal-to-noise
414: ratio of $(\sigma_n - \sigma_i)$. {\bf (c)} Linewidth ratio $(\sigma_i /
415: \sigma_n)$
416: for S/N of $(\sigma_n - \sigma_i) >$ 2.5. }
417: \end{figure}
418:
419: \begin{figure} % figure 4
420: \epsscale{0.6}
421: \plotone{f4.ps}
422: \caption[]{Magnetic Field structure in DR21(OH) MM1 core. The greyscale
423: shows the inclination angle of the magnetic field to
424: the line of sight ($\alpha$) varying from 30\deg\ to
425: 42\deg. The vectors represent
426: $B_p$ orientation inferred from dust (black) and CO(white) polarization.
427: $\oplus$ represnts the CN Zeeman measurement of line-of-sight component,
428: $\sim 0.36$mG towards observer. }
429: \end{figure}
430: \end{document}
431:
432: