astro-ph0109517/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: \usepackage[]{natbib}
3:  
4: \newcommand{\cthead}[1]{\multicolumn{1}{c}{#1}}
5: \newcommand{\ks}{km~s$^{-1}$}
6: \newcommand{\kss}{km~s$^{-1}$ }
7: \hyphenation{OH}
8: 
9: 
10: \newcommand{\vdag}{(v)^\dagger}
11: \newcommand{\myemail}{vslysh@asc.rssi.ru}
12: 
13: %% You can insert a short comment on the title page using the command below.
14: 
15: \slugcomment{Not to appear in Nonlearned J., 45.}
16: 
17: %% If you wish, you may supply running head information, although
18: %% this information may be modified by the editorial offices.
19: %% The left head contains a list of authors,
20: %% usually a maximum of three (otherwise use et al.).  The right
21: %% head is a modified title of up to roughly 44 characters.  Running heads
22: %% will not print in the manuscript style.
23: 
24: \shorttitle{Total linear polarization in the OH maser W75N}
25: \shortauthors{Slysh et al.}
26: 
27: %% This is the end of the preamble.  Indicate the beginning of the
28: %% paper itself with \begin{document}.
29: 
30: \begin{document}
31: 
32: %% LaTeX will automatically break titles if they run longer than
33: %% one line. However, you may use \\ to force a line break if
34: %% you desire.
35: 
36: \title{Total linear polarization in the OH maser W75N:\\
37: VLBA polarization structure.}  
38: 
39: 
40: %% Use \author, \affil, and the \and command to format
41: %% author and affiliation information.
42: %% Note that \email has replaced the old \authoremail command
43: %% from AASTeX v4.0. You can use \email to mark an email address
44: %% anywhere in the paper, not just in the front matter.
45: %% As in the title, you can use \\ to force line breaks.
46: 
47: \author{V. I. Slysh}
48: \affil{Astro Space Center, Lebedev Physical Institute, Profsoyuznaya
49:   84/32, 117810 Moscow, Russia}
50: \email{vslysh@asc.rssi.ru}  
51: 
52: \author{V. Migenes}
53: \affil{University of Guanajuato, Department of Astronomy, Apdo Postal 144, Guanojuato,
54:   CP 36000, GTO Mexico}
55: %\email{vmigenes@astro.ugto.mx}
56: 
57: \author{I. E. Val'tts}
58: \affil{Astro Space Center, Lebedev Physical Institute, Profsoyuznaya
59:   84/32, 117810 Moscow, Russia}
60: 
61: \author{S. Yu. Lyubchenko}
62: \affil{Astro Space Center, Lebedev Physical Institute, Profsoyuznaya
63:   84/32, 117810 Moscow, Russia}
64:   
65: \author{S. Horiuchi}
66: \affil{Astro Space Center, Lebedev Physical Institute, Profsoyuznaya
67:   84/32, 117810 Moscow, Russia}
68:   
69: \author{V. I. Altunin}
70: \affil{Jet Propulsion Laboratory, 4800 Oak Grove Dr.,
71: 	Pasadena, CA 91109, USA}
72:   
73: \author{E. B. Fomalont}
74: \affil{National Radio Astronomy Observatory, 520 Edgemont Rd., Charlottesville,      
75: 	VA22903, USA}
76: 
77: \and
78: 
79: \author{M. Inoue}
80: \affil{Nobeyama Radio Observatory, Nobeyama, Minamisaku,
81: 	Nagano 384-13, Japan} 
82: 
83: %% Notice that each of these authors has alternate affiliations, which
84: %% are identified by the \altaffilmark after each name.  Specify alternate
85: %% affiliation information with \altaffiltext, with one command per each
86: %% affiliation.
87: 
88: 
89: \begin{abstract}
90: W75N is a star-forming region containing various ultracompact HII regions 
91: and OH, H$_2$O, and CH$_3$OH maser emission. Our VLBA map shows that the OH 
92: masers are located in a thin disk rotating around
93: an O-star which is the exciting star of the ultracompact HII region VLA1.
94: A separate set of maser spots is connected with 
95: the ultracompact HII region VLA2. The radial velocity of OH maser spots
96: varies across the disk from 3.7~km~s$^{-1}$ to 10.9~km~s$^{-1}$. 
97: The diameter of the
98: disk is 4000~A.U. All maser spots are strongly polarized. This are the first
99: OH masers showing nearly 100 per cent linear polarization in several spots. 
100: Two maser spots seem to be Zeeman pairs corresponding to a magnetic field of 
101: 5.2~mgauss and 7.7~mgauss, and in one case we tentatively found a Zeeman pair consisting
102: of two linearly polarized components. 
103: The linearly polarized maser spots are shown to be 
104: $\sigma$-components which is 
105: the case of the magnetic field being perpendicular to the line of sight. 
106: The direction of the magnetic field as determined from  
107: linearly polarized spots is perpendicular to the plane of the disk, although
108: the galactic Faraday rotation may significantly affect this conclusion.
109: \end{abstract}
110: 
111: \keywords{ISM: radio lines: ISM -- masers -- surveys -- ISM: molecules}
112: 
113: 
114: \section{Introduction}
115: 
116: OH maser emission is strongly polarized, usually in a circular mode,
117: although an admixture of linear polarization was found in some masers,
118: which resulted in the elliptical polarization mode. The importance of 
119: the polarization properties of OH masers lies in the physics of the 
120: maser excitation and saturation properties. Current theories of OH 
121: maser emission polarization attribute it to the magnetic field in the 
122: emission region, of the order of several milligauss. This is three 
123: orders of magnitude larger than the general galactic magnetic field, 
124: but is consistent with estimates of the maser emission region density 
125: of $10^6-10^7$ ~cm$^{-3}$ as originating from the compression of the 
126: interstellar matter and magnetic field. Determination of the magnitude
127: and direction of the magnetic field will help to better understand the role
128: of the magnetic field in the formation of circumstellar disks or shocks
129: at the interface between compact HII regions and the ambient molecular cloud.
130: 
131: The polarization properties of OH maser emission were predicted by 
132: theoretical models on the basis of Zeeman splitting \citep*{goldreich73}.
133: If the Zeeman splitting exceeds the line width
134: (B$>$0.5 milligauss), three 100 per cent polarized components are predicted:
135: unshifted $\pi$--component linearly polarized along the projected 
136: direction of the magnetic field, and two 100 per cent elliptically polarized 
137: $\sigma$--components shifted in frequency to both sides of the $\pi$--component.
138: The relative intensity of the $\pi$--component and the degree of 
139: ellipticity depend on the angle $\theta$ between the line of sight and
140: the magnetic field direction. If $\theta=\pi/2$, the $\pi$--component 
141: is at maximum, and $\sigma$--components are linearly polarized 
142: perpendicular to the magnetic field direction. When $\theta=0$, the 
143: intensity of the $\pi$--component is  zero, and $\sigma$--components 
144: are 100 per cent circularly polarized in opposite directions. When 
145: $0<\theta<\pi/2$, the $\pi$--component is linearly polarized along the
146: magnetic field direction and is of intermediate intensity, and 
147: $\sigma$--components are elliptically polarized, with the 
148: major axis of the ellipsis perpendicular to the magnetic field. Such 
149: a picture has never been observed in OH masers in full details.
150: Typically, individual maser spots were found to be highly circularly 
151: polarized only in one sense. In some cases opposite circular polarization
152: was found to come from the same maser spot and was interpreted as a 
153: Zeeman pair. Linear polarization was rarely observed, and usually was
154: found as part of the elliptically polarized emission. The best-studied
155: region of OH maser polarized emission is W3(OH). \citet{garcia88}
156: found several dozens of circularly polarized components, with 16 being
157: elliptically polarized. Only three components had linear polarization
158: greater than circular polarization, with a maximum fractional linear
159: polarization of 46 per cent. Five Zeeman pairs were identified. 
160: \citet{garcia88} came to the conclusion that no features were detected that might
161: have been identified as $\pi$--components. Most of the details were 
162: single, circularly polarized features. In a similar polarization study
163: of G35.02$-$0.74N with a lower angular resolution \citet{hutawarakorn99}
164: found 25 circularly polarized spectral features, with four
165: Zeeman pairs and five features in which linear polarization was also 
166: present, with a maximum fractional polarization of 63 per cent. They were 
167: considered as $\sigma$--components. 
168: 
169: W75N is another well-studied OH maser source. It is associated with a 
170: star--forming region and an ultracompact HII-region, at a distance 2~kpc 
171: \citep{habing74}.
172: VLBI maps of the OH main line, 1665~MHz, show nine components
173: in the elongated region 1.5$''$ in extent \citep{haschick81}. From 
174: a single antenna polarization study \citet{haschick81} were able to 
175: determine polarization of some maser spots in their VLBI map, and 
176: suggested two Zeeman pairs with opposite circular polarization. A true
177: high angular resolution polarization study of W75N was conducted with
178: MERLIN by \citet{baart86}. They also found several Zeeman pairs 
179: of oppositely circularly polarized spectral features coming from the 
180: same position within measurement errors. In addition they suggested that 
181: there may be seven linearly polarized components present. This suggestion
182: was based on the presence of opposite circularly polarized pairs of 
183: spectral features coming from the same position and having the same 
184: radial velocity, in contrast with Zeeman pairs having different radial
185: velocities. No cross-correlation measurements were done in this study,
186: which are needed to measure linear polarization.
187: 
188: The prevailance of circular polarization and a scarcity of linear polarization
189: is a long standing problem in maser polarization theory.
190: \citet*{goldreich73} have proposed that the Faraday rotation in the emission 
191: region might destroy linear polarization, and only circular polarization
192: remains. If the linear polarization suggested by \citet{baart86} is 
193: real, it might be used for testing polarization models. The polarization 
194: structure of OH masers may be smeared by the lack of spectral and 
195: angular resolution when several adjacent features are mixed. \citet{baart86}
196: made their study with the angular resolution of 280~mas and 
197: spectral resolution of 0.3~km\,s$^{-1}$, which is insufficient as will be 
198: seen from our results.
199: 
200: In this paper we present a new polarization study of W75N with much
201: higher angular and spectral resolution, and  with a full polarization 
202: analysis making use of all Stokes parameters. This made it possible 
203: to exclude any contamination of spectral features by a contribution 
204: from other spectral features. In addition, for the first time, polarization
205: mapping of W75N in another main line OH transition, 1667~MHz, has been 
206: performed. 
207:           
208: \section{Observations and data reduction}
209: 
210: The observations were made on 1998 July 1 as a part of the survey of compact
211: OH masers suitable for further observations with the space ground 
212: interferometer \citep{migenes01}. We used the VLBA in a snap-shot mode, 
213: observing every source with a 
214: 5 minute scan. The synthesized beam of the array at the OH 
215: frequency, 1665/1667 MHz, was elliptical, 12$\times$4~mas 
216: with the position angle of the major axis $-$5$^\circ$,
217: which corresponds to the size and orientation of the VLBA at the
218: moment of the snap-shot observation.
219: The observations were 
220: carried out at all four OH line frequencies, in two circular 
221: polarizations, with the bandwidth of 125 kHz (25~km\,s$^{-1}$) divided in 
222: 128 spectral channels at each OH frequency which provided a spectral resolution
223: of 0.176~km\,s$^{-1}$. The correlation was done with the NRAO VLBA
224: correlator in Socorro in the full cross-correlation mode providing 
225: four correlated complex spectra RR, LL, RL and LR for each pair of 
226: antennas. These outputs can be used for determination of Stokes 
227: parameters in a standard way: 
228: 
229: 
230: \[
231: I=\frac{1}{2}(RR+LL)\hspace{15mm}
232: V=\frac{1}{2}(RR-LL)
233: \]
234: \begin{equation}
235: Q=\frac{1}{2}(RL+LR)\hspace{15mm}  
236: U=\frac{i}{2}(LR-RL)
237: \end{equation}
238: 
239: The amplitude calibration was performed using the system noise temperature 
240: and gain curves provided by VLBA Operations. Continuum sources 
241: 3C273 and OQ208 were used for the band pass and polarization 
242: calibration. Fig.~1a,b show images of 3C273 and OQ208 in I (contours) 
243: and linearly polarized intensity (vectors). The vectors
244: shown in Fig.~1a correspond to a fractional linear polarization
245: $m_L$=1.4 per cent;
246: the circular polarization measured on 3C273 is $m_C$=1.8 per cent.
247: No calibration of the position angle of the linear polarization
248: has been done. In principle, one could calibrate the position angle by 
249: rotating the measured linear polarization vector of 3C273 to the 
250: published value (e.g. \citep{conway93}) and applying this correction
251: to W75N OH maser data. However, we considered this to be of little
252: usefullness because observations of W75N have been performed 7.5 hours
253: after observations of 3C273, and during that time interval ionospheric
254: Faraday rotation could change significantly direction of the polarization
255: vector. Moreover, in view of the uncertain correction for the Galactic
256: Faraday rotation (Section 4.2) the absolute position angle of the linear
257: polarization remains unknown, and in what follows we give only relative
258: position angles which are rotated by an unknown amount. 
259: For OQ~208 linear polarization vectors correspond to $m_L\leq$0.1 
260: per cent they are distributed over the entire  field and  
261: are caused by noise; the circular polarization is also $m_C$=1.8 per cent.
262: OQ208 is known to be unpolarized \citep{stanghelini98}, while
263: the 18~cm linear polarization of the core of 3C273
264: shown on Fig~1a is $m_L$=1.8 -- 2.15 per cent \citep{conway93},
265: which is close to our result on Fig.~1a taking into account the crudeness
266: of the snap-shot measurements. In any case it would be safe to disregard
267: any linear polarization below 1 per cent. As for the circular polarization 
268: which 
269: is known to be absent in these two sources, the consistency of the two 
270: measurements means that there is unaccounted instrumental polarization 
271: in a sense that the intensity of the right circular polarization is 
272: systematically larger than the intensity of the left circular polarization.
273: In presenting the OH maser data we made a correction for the instrumental 
274: polarization by subtracting 1.8 per cent from all fractional polarization 
275: values of $m_C$. These corrections are small compared to the fractional 
276: polarization measured for OH spectral features.
277: 
278: The calibration was performed by the self-calibration of the reference
279: spectral feature using AIPS task "FRING". The reference feature was chosen 
280: such as to allow calibration of both RCP and LCP channels of the receiver.
281: This is possible if a spectral feature emits strongly enough in both modes
282: of polarization from the same position. To satisfy these requirements 
283: the feature must be either unpolarized, or linearly polarized. Since for OH 
284: masers
285: unpolarized features are rare, we looked for linearly polarized
286: features. Fig.~2 (upper) shows  the cross-polarized spectrum RL on the 
287: baseline 
288: Kitt Peak -- Pie Town. The non-zero features in this cross-polarization 
289: spectrum (i.e. a correlated spectrum from a pair of antennas with opposite
290: circular polarizations) are due to the linear polarization. There are several
291: linearly polarized features in the spectrum which can be taken as a reference.
292: We took the feature at 5.65~km\,s$^{-1}$ which showed the largest correlated 
293: flux on the longest baseline, Mauna Kea -- Saint Croix, and could be used for 
294: calibrating all the antennas of the array.
295: 
296: The calibration of the array was performed by means of fringe fitting
297: to the reference feature. A correction to the position of the phase center
298: adopted at the correlator was made using the measured absolute position of 
299: the reference feature. The absolute position was measured by the 
300: fringe rate method. Since the position offset of different spectral features
301: in W75N is known to be quite large compared to the size of the map, the 
302: mapping was carried out in two steps. First, an approximate position of all 
303: spectral features relative to the reference feature was determined with the 
304: fringe rate method. Then maps of spectral features were constructed centered
305: on these approximate positions. From the maps relative positions and angular
306: dimensions were determined by fitting two-dimensional Gaussians and 
307: deconvolving them with the beam. Separate maps of spectral features were 
308: obtained in all Stokes parameters: I, Q, U and V. From the first three
309: Stokes parameters polarization maps were constructed, which show total 
310: intensity I contours and linear polarization vectors. 
311: 
312: The 1667~MHz data were considered as independent from the 1665~MHz data and 
313: were calibrated separately, using the feature at the radial velocity of 
314: 9.8~km\,s$^{-1}$ (Fig.2 lower) as a reference. The absolute position of the 
315: 1667~MHz reference feature was
316: determined by the fringe rate method independently from the 1665~MHz 
317: measurements, with a lower accuracy than the relative position measurement 
318: accuracy.
319: 
320: \section{Results}
321: 
322: \subsection{Spectrum}
323: 
324: 
325: The cross-correlated circularly polarized spectra of the 1665 MHz line are 
326: shown in Fig.~3.  If compared with spectra taken in 1975 by \citet{davies77}
327: one can see many changes. The strongest component is the 
328: RCP feature near 12~km\,s$^{-1}$ with the same flux density, but with the 
329: radial velocity shifted to 12.45~km\,s$^{-1}$. A nearby weaker component at 
330: 13~km\,s$^{-1}$  has disappeared since 1975. Two new RCP features have 
331: appeared, at 0.65~km\,s$^{-1}$ and 3.0~km\,s$^{-1}$. In the LCP spectrum a new 
332: component at 9.4~km\,s$^{-1}$ has appeared, as well as a component at 
333: 0.65~km\,s$^{-1}$. The 1667~MHz spectrum shows
334: less spectral features, and they occupy a smaller velocity range. The total 
335: intensity cross-correlated spectra at 1665~MHz and 1667~MHz are shown in 
336: Fig. 4a,b. There seems to be no one-to-one correspondence between the 
337: spectral features in the two main OH lines. 
338: No emission features were found in 
339: the OH satellite lines at 1612~MHz and 1720~MHz.
340: 
341: \subsection{Absolute position}
342: 
343: The absolute position of the 1665~MHz reference feature is given in the 
344: footnote to Table~1. It is a factor of 3 more accurate than the previously 
345: measured absolute position \citep{haschick81}, but both measurements give 
346: coinciding positions within the combined error box. The absolute position of 
347: the strongest 1667~MHz feature at 9.8~km\,s$^{-1}$ was measured independently, 
348:  and was found to be shifted in the sky from the 1665~MHz reference feature 
349: by -- 175$\pm$69~mas in right acsension (offsets in right ascension mean 
350: R.A.$\times$cos(Dec)) and by -- 661$\pm$42~mas in declination.
351: 
352: \subsection{Distribution of maser spots}
353: 
354: Eleven maser spots were identified at the 1665~MHz transition and four at 
355: the 1667~MHz. In Table~1 we present the relevant parameters for the maser 
356: spots determined from their spectra and from polarization maps: LSR radial 
357: velocity, Stokes parameters
358: (I, V, Q, U), percentage of circular ($m_C$) and linear ($m_L$)
359: polarization, linear polarization position angle $\chi$, offsets in R.A. 
360: and Dec. from the reference position, deconvolved angular dimensions of
361: the fitted elliptical Gaussians and their position angle $\Theta$. The error in the 
362: relative positions is less than 1~mas. The position error of the 
363: 1667~MHz spot L is much larger since it was determined by fringe rate method,
364: but the position errors of the rest of the 1667~MHz spots relative to spot L
365: is less than 1~mas. 
366: The position of the maser spots is shown on Fig.~5. Also shown are the 
367: ultracompact HII~regions VLA1 and VLA2 \citep{torrelles97}. Most of the 
368: spots lie on an almost straight line of 1.6~$''$ (3200 A.U.) in extent, 
369: slightly 
370: displaced from VLA1. The 1667~MHz spots do not coincide with any of 1665~MHz
371: spots, however they are located in the same general area. The radial velocity 
372: varies along the line  from 3.8~km\,s$^{-1}$ to 10.9~km\,s$^{-1}$ 
373: (this is the mean radial velocity of features A and B, which come from the
374: same position as a Zeeman pair), moving from North to South. Two 
375: separate spots, J and K are located near the second ultracompact HII~region 
376: VLA2.  The map closely resembles the MERLIN map by \citet{baart86}. All 
377: our 1665~MHz spots can be found on the \citet{baart86} map, except for 
378: spot K at 0.65~km\,s$^{-1}$. There are several weak spots on the MERLIN map 
379: which are not present on our map but this can be due to our poorer 
380: sensitivity.  The 1667~MHz map has been built for the first time, and no 
381: comparison with earlier maps is possible. 
382: Also, a good correspondence is found with the VLBI map by \citet{haschick81}.
383: The only significant difference between our map and those of  
384: \citet{baart86} and \citet{haschick81} is the shifted position of spot 
385: A (12.45~km\,s$^{-1}$) by 140~mas; note, that the radial velocity of 
386: this feature has also changed by 0.45~km\,s$^{-1}$. One can not be sure that 
387: feature A is identical to the 12~km\,s$^{-1}$ feature in earlier maps, it is 
388: possible that the 12~km\,s$^{-1}$ feature has disappeared, and that feature A 
389: is a new feature which has flared up not far from the 12~km\,s$^{-1}$ 
390: position.  On the other hand the identification of feature A with the 
391: 12~km\,s$^{-1}$ feature is supported 
392: by its dominance in both spectra and by the equality of flux densities. 
393:             
394: \subsection{Spot size and brightness temperature}
395: 
396: All maser spots have been partially resolved. From Table~1 one can see that 
397: the largest spots are spot K (20~mas$\times$ 12~mas) and spot 
398: N (14~mas$\times$ 3~mas), but they may consist of several 
399: smaller spots which we can not resolve. 
400: The rest of the spots can be fitted with Gaussians with the major axis from
401: 4.4 to 10.7~mas, and the minor axis from 0.7  to 4.6~mas. 
402: Considering that the beam size is 12 $\times$ 4~mas, 0.7~mas 
403: must be regarded only as a formal fitting solution. The real extent of the 
404: minor axis can be much less than the indicated, so it should be considered 
405: as an upper limit. The brightness temperature of one of the brightest 
406: 1665~MHz spots, F (reference feature), with the integrated flux density 
407: of 18.1~Jy is $1.6\times$10$^{12}$K. The brightest 1667~MHz spot is L, with 
408: the integrated flux density 31.9~Jy, and it has a brightness temperature of 
409: 1.9$\times$10$^{12}$~K. These must be regarded as lower limits since
410: most of the spots are probably not resolved along the minor axis. 
411: All the mapped spots are elongated, with the major to minor axis ratio 
412: 3 or higher. 
413: If the size of the maser spots is not
414: intrinsic and is caused by the scattering in the interstellar medium, then
415: the elongation of the spots may be a result of the anisotropic scattering.
416: 
417: \subsection{Polarization of maser spots}
418: 
419: All maser spots given in Table~1 were imaged in all Stokes parameters.
420: This made it possible to determine full polarization properties of the maser 
421: spots. The percentage of circular polarization given in Table~1 was 
422: calculated as 
423: 
424: \begin{equation}   
425: \qquad m_C=100\frac{V}{I}
426: \end{equation}   
427: Positive $V$ corresponds to the right circular polarization. Percentage 
428: of the linear polarization was calculated as:
429: 
430: \begin{equation}   
431: \qquad m_L=100\frac{\sqrt{Q^2+U^2}}{I}                  
432: \end{equation}   
433: and the position angle of the electric vector of linear polarization was
434: calculated as 
435: 
436: \begin{equation}   
437: \qquad \chi=\frac{1}{2}\arctan\frac{U}{Q}
438: \end{equation}   
439: with positive direction East of North.
440: $I$, $Q$, $U$, $V$ are Stokes parameters at the maximum of the corresponding 
441: spot maps (Jy/beam). In some cases the percentage of the polarization exceeds
442: 100. This may be due to a difference in the position of maxima in different 
443: Stokes parameters, as well as to errors of intensity measurements. 
444: 
445: From Table~1 it is evident that the maser spots are highly polarized,
446: typically 100 per cent polarized. There are 2~types of the maser spots. The 
447: first type is strongly linearly polarized spots with $m_L>$70 per cent
448: (E,F,G,H,J,K). Polarization of the spots in this group is not purely 
449: linear, it has an admixture of small circular polarization, making the 
450: polarization elliptical. The percentage of circular polarization varies 
451: from less than 0.3 per cent to 80 per cent. \citet{baart86} suggested 
452: several candidates as linearly polarized components, based on the small 
453: difference
454: in position and equal velocities of several oppositely circularly 
455: polarized pairs. Now, with the full polarization analysis and a much 
456: higher position accuracy we can confirm or reject  \citet{baart86}
457: candidates. We confirm five of seven MERLIN candidates: J=A,A; G=F,C;
458: F=G,D; E=H,E; A,B=N,K. The former is the spot designation in Table~1,
459: and the latter are designations by Baart et al. (1986). Two other 
460: candidates, D,B and M,J were not present in our data. The spot A,B=N,K is
461: predominantly circularly polarized, with linear polarization of only 9.3 
462: per cent.  There is a pair of oppositely polarized spectral features at the
463: radial velocity 9.35~km\,s$^{-1}$, which looks like a linearly polarized 
464: feature suitable for the calibration; but it is absent in the cross-polarized 
465: spectrum in Fig.2a which means that the left-hand and right-hand polarized 
466: emission come from different positions, which has been confirmed by the 
467: subsequent mapping (features B and C in table 1). This is a chance
468: coincidence of radial velocities of two independent maser spots with opposite
469: circular polarization which could be mixed with linear polarization in the 
470: absence of the full polarization analysis and high angular resolution 
471: mapping. In the 1667~MHz transition there is only one linearly polarized 
472: spot, with a percentage of linear polarization of 41.6 per cent, and with 
473: a circular polarization fraction of 23 per cent. 
474: 
475: The position angle of the linear polarization vector is in the range of 
476: $-23^\circ$ to $23^\circ$ for the spots in this group. The mean 
477: position angle of the linear polarization vector is $3.9^\circ\pm17^\circ$ 
478: (Fig.~6). Fig.~7 shows polarized and total intensity images
479: of several linearly polarized spots. 
480: 
481: The other type of maser spots is strongly circularly polarized, with an
482: admixture of linear polarization (spots A--D,I,M,N,O). The percentage of 
483: the circular polarization in this group is close to 100 per cent, and the 
484: fractional linear polarization is of less than 5 per cent to 17.9 per cent, 
485: and only spot L has moderate linear and circular polarization.
486: 
487: \subsection{Identification of Zeeman pairs}
488: 
489: The 1665~MHZ emission from spots A and B, and from spots M and O at 1667~MHz 
490: come from the same positions, with the position difference 0.8~mas 
491: and 1.1~mas, respectively, which is much less than the size of the 
492: beam and the size of the spots. 
493: The polarization is oppositely circular, and the velocity difference
494: is 3.07~km\,s$^{-1}$ and 2.7~km\,s$^{-1}$, respectively.  Based on these 
495: properties one can identify these two pairs as $\sigma{\pm}$-components of 
496: the Zeeman pattern, with the magnetic field 5.2~mgauss and 7.7~mgauss, 
497: respectively (positive sign indicates field direction pointing away from the 
498: Earth). Both pairs are elliptically polarized, with nearly the same position
499: angles of the polarization vector, exactly as required for 
500: $\sigma$--components. Nevertheless there are distinctions from the theoretical
501: Zeeman pattern. The first is the inequality of the intensities of right-hand 
502: and left-hand polarized emission. The intensity ratio of circularly polarized 
503: components is 5 for the A,B pair, and 1.8 for the M,O pair. Linearly 
504: polarized intensities are also different, the ratio being 2.2 for the A,B 
505: pair and 1.2 for the M,O pair.
506: Another important deviation from the theoretical Zeeman pattern is the  
507: complete absence of $\pi$--components. In a Zeeman pair the $\pi$--component
508: must be located at the mean radial velocity of the pair. We conducted an 
509: intense search for $\pi$--components for the two Zeeman pairs: A,B at the 
510: radial velocity of 10.9~km\,s$^{-1}$ and for M,O at the radial velocity of 
511: 7.3~km\,s$^{-1}$, and did not find any emission above the noise level. Also, 
512: a $\pi$--component must be only linearly polarized, and we found that every 
513: maser spot is elliptically polarized. The only spot for which the circular 
514: polarization was not detected is spot E. Therefore we conclude that 
515: $\pi$--components are not present in the Zeeman spectrum of OH masers in W75N. 
516: \citet{baart86} suggested several Zeeman pairs, none of which are  
517: confirmed here. For the pair with radial velocities of 3.8~km\,s$^{-1}$ (LCP) 
518: and 6.5~km\,s$^{-1}$ (RCP) suggested by  \citet{baart86} we confirm only 
519: the LCP-component, with LCP/RCP ratio above 28; the LCP/RCP ratio in  \citet{baart86}
520: is 5. In addition, for the other Zeeman pair suggested by  \citet{baart86}
521: with radial velocity of 4.8~km\,s$^{-1}$ (LCP) and 
522: 9.6~km\,s$^{-1}$ (RCP) we could not find the LCP counterpart, with the 
523: RCP/LCP ratio of more than 350 while  \citet{baart86} data suggest 
524: RCP/LCP=12. 
525: 
526: \citet{haschick81} suggested two possible Zeeman pairs which are not
527: present in our data. The discrepancy between our data and earlier results 
528: can be attributed to the the variability of the Zeeman component intensity.
529: We never see equal intensity of $\sigma$--components, and the intensity
530: ratio can vary in a very large range. If the $\sigma$--component intensity 
531: ratio is very large, the weaker component becomes unobservable. This may  
532: explain the presence of single circularly polarized spectral features 
533: in OH masers.        
534: 
535: \subsection{Polarization status}
536: 
537: On Fig.~8 we show the percentage of circular polarization plotted against
538: the percentage of linear polarization for all spots. The spots are labelled
539: by letters as in Table~1. The dashed line shows where completely polarized
540: features could appear on the diagram. The majority of the spots are located
541: close to the 100 per cent polarization line. A notable exception is spot L 
542: (1667~MHz transition) which is 47 per cent polarized. There is evidence for  
543: a correlation in that predominantly circularly polarized spots are either 
544: located in the upper and lower left corners of the diagram, and predominantly 
545: linearly polarized spots are located in the middle right of the diagram. 
546: The numbers on the dashed line indicate an angle $\theta$ between the line 
547: of sight and the magnetic field direction, according to equations
548: 
549: \begin{equation}
550: m_L=100\frac{\sin^2\theta}{1+\cos^2\theta}\hspace{15mm}
551: m_C=\pm100\frac{\cos\theta}{1+\cos^2\theta}  
552: \end{equation}
553: 
554: These equations describe polarization of the $\sigma$--components in the 
555: theory by  \citet*{goldreich73} for the case when the Zeeman
556: splitting is larger than the linewidth  \citep[see also][]{elitzur96}. Most 
557: of the maser spots occupy two regions on the diagram: 
558: $\theta <30^\circ$ and $80^\circ<\theta<90^\circ$. If all the spots are 
559: $\sigma$--components it means that the line of sight is either parallel or 
560: perpendicular to the magnetic field direction, and no intermediate cases exist.
561: The predominantly circularly polarized spots are almost certainly $\sigma$--components,
562: although frequently without the second member of the pair.
563: The predominantly linearly polarized spots are also $\sigma$--components 
564: with $\theta\approx$90$^\circ$, or $\pi$--components. A correct attribution is 
565: important for the determination of the direction of the magnetic field:
566: it is parallel to E-vector for $\pi$--components, and perpendicular to 
567: E-vector for $\sigma$--components. As mentioned in Section 3.6 the presence 
568: of the small circular polarization makes these components elliptically 
569: polarized, and therefore they must be $\sigma$--components, since 
570: $\pi$--components can not have a circular polarization contribution. 
571: 
572: The only linearly polarized spot, L at 1667~MHz, is definitely not 100 per 
573: cent polarized and is an exception among all maser spots in W75N. In the model
574: of the weak magnetic field the percentage of linear polarization for 
575: $\sin^2\theta>1/3$ is 
576: 
577: \begin{equation}
578: m_L=100\frac{3\sin^2\theta-2}{3\sin^2\theta} 
579: \end{equation}
580: 
581: \citep*{goldreich73}. For this spot $m_L= 41.6$ per cent, and 
582: corresponds to $\theta=43.4^\circ$. The percentage of circular polarization 
583: varies across the line profile (Fig.~9) changing the sign as predicted by the
584: theory for the case of small Zeeman splitting \citep{elitzur96}. However the curve
585: is shifted from the line centre and is strongly distorted compared to the 
586: theoretical one, probably because of the non-linear mode competition in 
587: the maser. The linear polarization increases to the red-side of the profile,
588: while the circular polarization is larger at the blue side.
589: 
590: \section{Discussion}
591: 
592: \subsection{Linear polarization}
593: 
594: The most important result of this study is the discovery of several linearly
595: polarized maser spots with almost 100 per cent polarization. The mere 
596: existence 
597: of linear polarization in these spots and in the circularly polarized
598: spots means that the Faraday rotation proposed by \citet*{goldreich73}
599: as a mechanism for the elimination of $\pi$--components is 
600: not important in W75N OH maser. The linearly polarized components are
601: $\sigma$--components of the Zeeman pattern.
602: This conclusion can be tested if the second $\sigma$--component of the Zeeman pair of linearly 
603: polarized emission
604: will be found, shifted in radial velocity by an appropriate amount. We tentatively found
605: the Zeeman counterpart F1 to the linearly polarized feature F. 
606: It is shifted in the radial velocity by
607: 6.3~km\,s$^{-1}$, corresponding to the magnetic field strength 10.7 milligauss, and
608: is almost fully linearly polarized, with the position angle close to the
609: position angle of spot F. This is consistent with the $\sigma$--component interpretation of spots
610: F and F1. However, this conclusion has to be confirmed with a more sensitive measurements, 
611: since
612: the detection of the component F1 is tentative. Its intensity is only about 0.003 of intensity of
613: component F, and it was detected on the edge of the spectral band of the receiver, in fact, in
614: the last 128th channel. This is not surprising since the Zeeman pairs were found in W75N, with
615: unequal intensity of the components. The linearly polarized Zeeman counterpairs
616: might be too weak for a detection. The difference in intensity of Zeeman $\sigma$--components
617: seems to be a widespread property of OH masers, and is probably due to a gradient of the radial
618: velocity and magnetic field \citep{cook66}. The absence of $\pi$--component is more difficult to 
619: understand.
620: As mentioned above internal Faraday rotation can not be responsible for this.
621: Maser model calculation by \citet*{gray95} produce both $\pi$  and $\sigma$--components, 
622: with
623: $\sigma$--components dominating over $\pi$--components in the range of angles
624: $\theta=0-55^\circ$, where $\sigma$--components are circularly or elliptically polarized. For
625: $\theta=55-90^\circ$, $\pi$--components which are linearly polarized dominate.
626: Saturation effects of competitive gain strongly reduce a weaker component, that is
627: $\pi$--component for $\theta=0-55^\circ$, and $\sigma$--component 
628: for $\theta=55-90^\circ$.
629: \citet*{gray95} note that the suppression of $\sigma$--component is much less
630: efficient, than the suppression of $\pi$ by $\sigma$--components.
631: The model of \citet*{gray95} is in conflict with our conclusion that
632: linearly polarized spots in W75N are emitting $\sigma$--components with $\theta$
633: close to $90^\circ$, while the model predicts dominance of the linearly polarized
634: $\pi$--components for $\theta$ near $90^\circ$. Our identification of the linearly
635: polarized spots as $\sigma$--components was based on the presence of some circularly
636: polarized emission in all spots except spot E. This seems to be a very firm
637: argument in favor of $\sigma$--components unless an explanation for the 
638: circular polarization contribution to the $\pi$--components is found.
639: 
640: There is an alternative interpretation of the observed linear polarization as a result
641: of the saturated maser amplification with a weak magnetic field, when
642: the Zeeman splitting is less than the bandwidth. For $\sin^2\theta\le1/3$,
643: or $\theta\le35.5^\circ$ there will be a single 100 per cent linearly polarized emission line
644: with E-vector parallel to the magnetic field direction (Case 2a in \citet*{goldreich73}).
645: For $\theta>35.5^\circ$ the linear polarization will be lower, and the 
646: circular polarization appears 
647: in the line wings (see Section 3.7). This model could explain observed linearly polarized emission in 
648: W75N
649: but the observed admixture of circular polarization remains unaccounted, the same as for
650: the $\pi$--components
651: in the Zeeman splitting. Also this model limits the magnetic field to about 0.5 milligauss which is 
652: much less
653: than was deduced for spots A, B, F, F1 and M, O in Section 3.6. More likely this explanation could 
654: be applied
655: to the observed linear polarization of H$_2$O and methanol masers. And, finally, the detection of the
656: Zeeman counterpart F1 the spot F, if confirmed, is a definite proof of the interpretation of the 
657: linearly
658: polarized emission features as $\sigma$--components of Zeeman pattern.
659: 
660: \subsection{The nature of maser spots}
661: 
662: We have mapped 14 maser spots in W75N (Fig. 5), with two or possibly three coinciding Zeeman 
663: components.
664: Every maser spot has its own radial velocity, different from the radial velocity of other spots, and a
665: narrow line width of about 0.2 -- 0.3~km\,s$^{-1}$. 1667-MHz spots do not coincide in position with 
666: 1665-MHz
667: spots. Almost every spot is 100 per cent elliptically polarized. We interpret them as $\sigma$--
668: components
669: of the Zeeman pattern. The magnetic field in the Zeeman pairs was determined to be from 5 to 11 
670: milligauss,
671: the same values as found in other OH masers \citep{garcia88}. One can assume that in 
672: other spots where the second
673: $\sigma$--component is not visible the magnetic field strength is of the same order.
674: If it is a $\sigma$--component, then the linear 
675: polarization
676: vector must be perpendicular to the direction of the magnetic field, and the
677: magnetic field  direction seems to be roughly the same for all spots.
678: This can be seen on the
679: histogram in Fig~6. Unfortunately the direction of the magnetic field 
680: can not be determined reliably from the linear polarization measurements
681: because of the uncertain amount of Faraday 
682: rotation in the Galaxy. The OH masers are located in the Galactic
683: disk where the Faraday rotation is at
684: maximum. Total Faraday rotation in the Galactic plane as measured
685: with extragalactic radio sources is
686: 
687: \begin{equation}
688: RM(l)=RM_0\sin(l_0-l) 
689: \end{equation}
690: 
691: where {\it RM$_0$}=1607~rad\,m$^{-2}$ $\pm$10~per cent r.m.s., {\it l$_0$}=62.1$^\circ$ \citep{clegg92}.
692: W75N is at the galactic longitude
693: {\it l}=81.9$^\circ$, and from (7) one has {\it RM}=--544~rad\,m$^{-2}$.
694: At the OH frequency ($\lambda=18$~cm) the Faraday rotation is (--544)$\times$(0.18)$^2$=--
695: 17.6~rad$\pm$5.4~rad if one assumes maximum uncertainty 30 per cent.
696: This is the total Faraday rotation throughout the galactic disk; W75N is at the distance of 2 kpc, which is
697: probably about a half of the total effective distance, and the Faraday rotation to W75N can be a
698: factor of 2 lower, or --9~rad$\pm$2.7~rad. 
699: This is a large rotation, about 3 full turns, and a correction
700: for the Faraday rotation to the position angle of the linear polarization could be quite uncertain.
701: Therefore it is not posible to determine the direction of the magnetic field in OH maser spots.
702: 
703: 	Physical parameters of the maser spots can be estimated from maser models, which require 
704: gas
705: density n$_{H_2}$=10$^7$cm$^{-3}$, kinetic temperature 100~K, dust temperature
706: 150~K, and OH abundance 10$^{-5}$ \citep*{gray95}. Such parameters can
707: provide inversion of 1665~MHz OH transition in a model with FIR line overlap and a velocity
708: gradient of about 0.025~km\,s$^{-1}$/A.U. \citep*{gray95}. With the magnetic
709: field strength of 10 milligauss the model of \citet*{gray95} provides 100 per cent
710: elliptically polarized $\sigma$--components, with $\pi$--components suppressed,
711: in agreement with results of this paper for W75N. The size of maser spots 10 A.U. and
712: molecular hydrogen density 10$^7$cm$^{-3}$ correspond to the mass of maser spots
713: of 2$\times10^{-7}$M$_{\odot}$ , which is less than the mass of the Earth.
714: 
715: If the maser spots are discrete physical objects -- dense cold gas condensations
716: surrounded  a low density medium -- they should be confined
717: by the external pressure. A gas condensation with the density 10$^{7}$cm$^{-3}$
718: and temperature 100~K can be in pressure equilibrium with the gas of density
719: 10$^{5}$cm$^{-3}$ and temperature 10$^4$~K.
720: However, the magnetic pressure in the maser spots, with the magnetic field
721: strength of 10 milligauss, is an order of magnitude higher, and can
722: be compensated by turbulent or ram pressure of the hot medium
723: (see discussion in a paper by \citet{reid87}).
724: Another model of maser spots proposed
725: for Class~II methanol masers \citep{slysh99} assumed that the maser spots 
726: are
727: extended gaseous envelopes of solid icy planets orbiting around O, B-stars, outside
728: their HII regions. OH molecules as well as methanol molecules are continuosly
729: supplied to the envelope by evaporation of ice from the surface of the planets.
730: In W75N the ultracompact HII region VLA1 (Fig.~5) marks the position of the central star
731: with luminosity 1.4$\times10^{5}$L$_{\odot}$ \citep{moore91} which corresponds to the
732: main sequence O9--star with mass 20M$_{\odot}$. The largest distance from VLA1 to
733: a maser spot is about 1000 mas, or 2000 astronomical units. At this distance from
734: a 20M$_{\odot}$ star the orbital velocity is 3~km\,s$^{-1}$ which is consistent with the observed
735: velocity range of maser spots in W75N. In this model the magnetic
736: field originates in the planets.
737: 
738: \subsection{The model of region W75N}
739: 
740: OH maser spots in W75N are located around the ultracompact HII region VLA1, 
741: with
742: spots J and K possibly associated with another ultracompact HII region, VLA2 
743: (Fig~5).
744: One model of OH masers places them at D-type ionization front
745: \citep*{elitzur78}
746: or in the photo-dissociation region \citep*{hartquist91} surrounding
747: the HII-region. In W75N the geometry of OH maser spots is consistent with these
748: models although the observed gradient of the radial velocity along the chain
749: of maser spots from about 10~km\,s$^{-1}$ at spots A, B to 3.7~km\,s$^{-1}$ at
750: spot I is not predicted by the models. This gradient is better represented in
751: the model of jet or in disk models. \citet{torrelles97} based on the elongated shape
752: of the ultracompact HII-region VLA1 and orientation of chains of H$_2$O and OH
753: maser spots parallel to the direction of the bipolar outflow, observed in W75N
754: on a much larger scale, suggested that the masers trace the outflow at scales
755: of about 1 arcsec and that VLA1 is the powering source of the molecular outflow.
756: Our more accurate measurements of the absolute position of OH masers show
757: that they are not projected at VLA1 (Fig~7 in \citet{torrelles97}) but are
758: displaced to form an arc arround VLA1. The H$_2$O masers seem to form an arc
759: which is closer to VLA1.  Such a geometry is better described by
760: a disk arround VLA1 in which OH maser spots are located at a distance
761: of about 2000 astronomical units from VLA1, and H$_2$O masers are located a factor
762: of 2 to 5 closer. The disk model was first suggested by \citet{haschick81}.
763: VLA1 marks the position of the ionization source and the center
764: of gravity of the system. As was suggested in Section 4.2 there
765: might be a 20M$_{\odot}$ O9-star at the
766: center of VLA1, which gravitationally holds the disk of radius 2000 astronomical units.
767: In this model the elongated shape of VLA1 reflects the distribution of disk material,
768: which became visible in radio due to the ionization by the central star. 
769: We can not determine direction of the magnetic field in the disk from linear polarization
770: data because of the  uncertain amount of Galactic Faraday rotation (see Section 4.2) but
771: we can state that the distribution of the magnetic field orientation is not random
772: and is well organized in some direction, as is evident from the
773: histogram of Fig~6.
774: 
775: \section{Conclusions}
776: \begin{itemize}
777: \item One of the results of this study is the discovery of several linearly polarized
778: maser spots with almost 100 per cent polarization.
779: 
780: \item For the first time polarization mapping of W75N in another main line OH transition,
781: 1667 MHz was done.
782: 
783: \item We suggest that the OH masers are located in a thin
784: disk rotating around an O-star
785: which is the exciting star of the ultracompact HII region VLA1.
786: The diameter of the disk is 4000 A.U. Two separate maser
787: spots are connected with the ultracompact
788: HII region VLA2.
789: 
790: \item Spots A and B at 1665~MHz and M and O at 1667~MHz come from the same
791: positions. The polarization is oppositely circular, and the position difference
792: is very small. Based on these properties one can identify these two pairs as
793: $\sigma{\pm}$--components of Zeeman patten. Both pairs are elliptically polarized,
794: with nearly the same position angles of the polarization vector, exactly as
795: required for $\sigma$--components. Two maser spots show Zeeman pairs corresponding
796: to the magnetic field 5.2 mgauss and 7.7 mgauss.
797: 
798: \item The linearly polarized maser spots are shown to be $\sigma$--components in
799: the case of the magnetic field perpendicular to the line of sight. We 
800: tentatively detected a Zeeman pair consisting of the linearly polarized 
801: components, with a deduced magnetic field of 10.7 mgauss. The direction of 
802: the magnetic field as determined for linearly polarized spots is perpendicular
803: to the plane of the disk, not corrected for the Galactic Faraday rotation.
804: \end{itemize}
805: 
806: \begin{center}
807: \bf{Acknowledgements}
808: \end{center}
809: 
810: NRAO is a facility of the National Science Foundation operated under a cooperative
811: agreement by Associated Universities. VIS and IEV acknowledge partial support
812: from INTAS (grant N97-11451) 
813: and Russian Foundation for Basic Research (grant N01-02-16902).
814: 
815: 
816: \begin{thebibliography}{}
817: \bibitem[Baart et al.(1986)]{baart86} Baart, E. E., Cohen, R. J., Davies, R. D., Norris, 
818: R. P., and Rowland, P. R. 1986, \mnras, 219, 145
819: \bibitem[Clegg et al.(1986)]{clegg92} Clegg, A. W., Cordes, J. M., Simonetti, J. K., and
820: Kulkarni, S. R.  1992, \apj, 386, 143
821: \bibitem[Conway et al.(1993)]{conway93} Conway, R. G., Garrington, S. T., Perley, 
822: R. A., and Biretta, J. A. 1993, \aap, 267, 347
823: \bibitem[Cook(1966)]{cook66} Cook, A. H., 1966 \nat, 211, 503 
824: \bibitem[Davies et al.(1977)]{davies77} Davies, R. D., Booth, R. S., and Perbet, J.-N.
825: 1977, \mnras, 181, 83 
826: \bibitem[Elitzur(1996)]{elitzur96} Elitzur, M. 1996, \apj, 457, 415
827: \bibitem[Elitzur and de Jong(1978)]{elitzur78} Elitzur, M., and de Jong, T. 1978, \aap, 67, 323 
828: \bibitem[Garcia-Barreto et al.(1988)]{garcia88} Garcia-Barreto, J. A., 
829: Burke, B. F., Reid, M. J., Moran, J. M., Haschick, A. D., and
830: Schilizzi, R. T. 1988, \apj, 326, 954  
831: \bibitem[Goldreich, Keeley, and Kwan(1973)]{goldreich73} Goldreich, P., Keeley, D. A., and Kwan, 
832: J. Y. 1973, \apj, 179, 111
833: \bibitem[Gray and Field(1995)]{gray95} Gray, M. D., and Field, D. 1995, \aap, 298, 243     
834: \bibitem[Habing et al.(1974)]{habing74} Habing, H. J., Goss, W. M., Mattews, H. E., and
835: Winnberg, A. 1974, \aap, 35, 1
836: \bibitem[Hartquist and Sternberg(1991)]{hartquist91} Hartquist, T. W., and Sternberg, A., 1991, 
837: \mnras, 248, 48      
838: \bibitem[Haschick et al.(1981)]{haschick81} Haschick, A. D., Reid, M. J., Burke, 
839: B. F., Moran, J. M., and Miller, G. 1981, \apj, 244, 76 
840: \bibitem[Hutawarakorn and Cohen(1999)]{hutawarakorn99}Hutawarakorn, B. and Cohen, R. J., 1999, \mnras, 303, 845 
841: \bibitem[Migenes et al.(2001)]{migenes01} Migenes, V., 
842: Slysh, V. I., Val'tts, I. E., Horiuchi, S., and Inoue, M. 2001, \apj, (in 
843: preparation)
844: \bibitem[Moore et al.(1991)]{moore91} Moore, T. J. T., 
845: Mountain, G. M., and Yamashita, T. 1991, \mnras, 248, 79
846: \bibitem[Reid et al.(1987)]{reid87} Reid, M. J., 
847: Myers, P. C., and Bieging, J. H. 1987, \apj, 312, 830
848: \bibitem[Slysh et al.(1999)]{slysh99} Slysh, V. I., Val'tts, I. E., Kalensky, S. V., and
849: Larionov, G. M. 
850: 1999, Astronomy Report, 43, 785
851: \bibitem[Stanghelini et al.(1998)]{stanghelini98} Stanghelini, C., Dallacasa, D., 
852: O'Dea, C. P., Baum, S. A., Fanti, C., and Fanti, R. 
853: 1998, Proceedings of the IAU Colloquium 164, eds. J. A. Zensus, G. B. Taylor,
854: \& J. M. Wrobel, A.S.P. Conference Series, p.177  
855: \bibitem[Torrelles et al.(1997)]{torrelles97} Torrelles, J. M., Gomez, J. F., Rodriguez, L. F., 
856: Ho, P. T. P., Curiel, S., and Vazquez, R.
857: 1997, \apj, 489, 744
858: \end{thebibliography}
859: 
860: \clearpage
861: 
862: %\begin{table*}
863: \begin{deluxetable}{lrrrrrrrrrrrr}
864: %\tablenum{1}
865: \tabletypesize{\scriptsize}
866: \rotate
867: \tablewidth{0pt}
868: \tablecaption{Maser spots in W75N}
869: \tablehead{
870: %&&&&&\nl
871: \colhead{Spot}&\colhead{v$_{lsr}$}&\colhead{$I$}&\colhead{$V$}&\colhead{$Q$}&\colhead{$U$}&\colhead{
872: $m_C$}&\colhead{$m_L$}&\colhead{$\chi$}&\colhead{$R.A.$}&\colhead{$Dec.$}&\colhead{Size}&\colhead{$\Theta$}\nl 
873: &\colhead{km\,s$^{-
874: 1}$}&\colhead{Jy/beam}&\colhead{Jy/beam}&\colhead{Jy/beam}&\colhead{Jy/beam}&\colhead{per 
875: cent}&\colhead{per 
876: cent}&\colhead{$^\circ$}&\multicolumn{2}{c}{mas}&\colhead{mas}&\colhead{$^\circ$} }
877: 
878: \startdata
879: \multicolumn{13}{c}{\bf{}1665 MHz}\nl
880: 
881: A & 12.45 & 16.88 &   16.24 & $<$0.04 & $-$1.56 &     94.4 &   9.3 & $-$45.0 & $-$301.7 & $-
882: $1176.7 &  5.9$\times$1.4 &  95\nl
883: B &  9.38 &  3.92 & $-$3.24 & $-$0.17 & $-$0.68 &  $-$84.4 &  17.9 & $-$52.2 & $-$301.1 & $-
884: $1177.3 &  5.5$\times$4.6 &  86\nl
885: C &  9.35 &  3.27 &    3.50 &    0.11 & $-$0.19 &    105.3 &   6.9 & $-$29.8 &  $-$98.2 &  $-$573.5 
886: &  5.6$\times$4.1 &  61\nl
887: D &  7.35 &  0.61 &    0.62 & $<$0.02 & $<$0.02 &    101.2 &$<$5.0 &    \dots&  $-$1.0 &   $-$32.7 
888: &  4.4$\times$1.1 &  83\nl
889: E &  6.00 &  5.76 & $<$0.02 &    4.71 &    2.69 &   $<$0.3 &  94.2 &    14.9 & $-$122.7 &  $-$663.6 
890: &  5.9$\times$2.2 &  92\nl
891: F $^{*}$ &  5.65 & 10.16 & $-$0.81 &    6.93 &    7.12 &   $-$9.8 &  97.8 &    22.9 &      0.0 & 0.0 &  
892: 5.4$\times$1.2 &  96\nl
893: G &  5.29 &  3.30 & $-$0.78 &    2.37 &    1.87 &  $-$25.4 &  91.4 &    19.2 &    155.5 &     114.2 &  
894: 7.6$\times$2.9 &  59\nl
895: H &  5.11 &  2.86 & $-$2.24 &    1.41 & $-$1.45 &  $-$80.0 &  70.7 & $-$22.9 &    186.9 &     167.1 
896: &  5.1$\times$2.5 &  94\nl
897: I &  3.70 &  1.18 & $-$1.28 & $<$0.02 &    0.13 & $-$110.8 &  10.8 &    40.6 &    249.0 &     317.9 
898: &  6.5$\times$2.9 &  77\nl
899: J &  3.00 &  8.16 &    0.79 &    7.61 & $-$3.05 &      7.9 & 100.5 & $-$10.9 &    299.8 & $-$1275.8 
900: & 10.7$\times$2.4 & 114\nl
901: K &  0.65 &  0.83 &    0.22 &    0.78 & $<$0.009&     24.2 &  93.7 &     0.0 &    376.5 & $-$1385.3 
902: & 19.7$\times$11.9& 142\nl
903: F1& -0.65 &  0.040 & $<0.003$ &  0.035 & 0.023&    $<10$ &  87   &    20.5 &     0.14 &      0.47 &  
904: 6.5$\times$0.7&   74\nl
905: \multicolumn{13}{c}{\bf{}1667 MHz}\nl
906: L & 9.80 &  21.33 & $-$4.46 &    6.01 &    6.54 &  $-$22.7 &  41.6 &    23.7 & $-$175.4 &  $-$661.2 
907: &  6.2$\times$1.9 &  91\nl
908: M & 8.65 &   1.49 &    1.26 & $<$0.01 &    0.14 &     82.5 &   9.4 &    45.0 & $-$149.9 &  $-$506.1 
909: &  5.4$\times$0.7 &  90\nl
910: N & 6.55 &   0.55 & $-$0.61 &    0.045& $<$0.01 & $-$112.8 &   8.5 &     8.0 & $-$136.2 &  $-
911: $526.8 & 14.2$\times$3.2 & 176\nl
912: O & 5.95 &   2.46 & $-$2.23 & $<$0.01 &    0.17 &  $-$92.5 &   7.1 &    45.0 & $-$149.7 &  $-
913: $505.0 &  7.4$\times$2.5 & 113\nl
914: 
915: \enddata
916: %\tableline
917: %\end{tabular}                       
918: \begin{list}{}{}
919: \item[$^*$]Reference position: R.A.=20$^{\rm h}$38$^{\rm m}$36$\fs$414$\pm$0$\fs$003 
920: Dec.=42$\degr$37$\arcmin$35$\farcs$44$\pm$0$\farcs$03 (J2000)
921: \end{list}
922: \end{deluxetable}
923:                         	       
924: \clearpage
925: 
926: \begin{figure}
927: \epsscale{0.30}
928: \plotone{f1a.eps}
929: \plotone{f1b.eps}
930: \caption{Continuum maps of calibration sources.
931: a) 3C273.
932: Contours - total intensity I: 0.9, 4.6, 8.3 Jy/beam;
933: vectors - linearly polarized intensity, maximum length=0.125 Jy/beam;
934: m$_L=1.4\%$; m$_C=1.8\%$ 
935: b) OQ208.
936: Contours - total intensity I: 0.085, 0.45, 0.77 Jy/beam;
937: vectors - linearly polarized intensity; 1 mas=0.0004 Jy/beam.
938: m$_L\le0.8\%$; m$_C=1.8\%$.  }
939: 
940: \end{figure}
941: 
942: \clearpage
943: 
944: \begin{figure}
945: \epsscale{0.80}
946: \plotone{f2.eps}
947: \caption{ The cross power spectra of W75N on the base line Kitt Peak - Pie Town.
948: Upper: cross-polarized RL spectrum of 1665~MHz OH line;
949: lower: cross-polarized RL spectrum of 1667~MHz OH line. }
950: 
951: \end{figure}
952: 
953: \clearpage
954: 
955: \begin{figure}
956: \epsscale{0.80}
957: \plotone{f3.eps}
958: \caption{The cross power circularly polarized spectra of W75N on the base line
959: Kitt Peak - Pie Town in 1665~MHz OH line.
960: Dotted line: left-circular polarization (LCP);
961: solid line: right-circular polarization (RCP).}
962: 
963: \end{figure}
964: 
965: \clearpage
966: 
967: \begin{figure}
968: \epsscale{0.80}
969: \plotone{f4.eps}
970: \caption{Total intensity I cross power spectra of W75N in 1665~MHz (upper) and
971: 1667~MHz (lower) OH lines.
972: Letters A - O mark spectral features which were mapped and their map and polarization
973: parameters are given in Table 1.} 
974: \end{figure}
975: 
976: \clearpage
977: 
978: \begin{figure}
979: \epsscale{0.80}
980: \plotone{f5.eps}
981: \caption{ A VLBA map of the OH maser features. Circles denote
982: 1665~MHz features, triangles - 1667~MHz features. They are labeled as in Table 1.
983: Vectors attached to the maser features show relative direction of the
984: magnetic field. An ellipse and a circle labelled VLA1 and VLA2 show position and size of ultra compact
985: HII regions, from Torrelles et al (1997). Dashed line shows a portion of a possible disk centered
986: at VLA1 and inclined by 12$^\circ$ to the line of sight.}
987: 
988: \end{figure}
989: 
990: \clearpage
991: 
992: \begin{figure}
993: \epsscale{0.80}
994: \plotone{f6.eps}
995: \caption{Distribution of the linear polarization relative position angles $\chi$ for maser spots
996: in W75N.  $\bar{\chi}=4^\circ\pm17^\circ$.  }
997: 
998: \end{figure}  
999: 
1000: \clearpage
1001: 
1002: \begin{figure}
1003: \epsscale{0.30}
1004: \plotone{f7a.eps}
1005: \plotone{f7b.eps}
1006: \plotone{f7c.eps}
1007: \plotone{f7d.eps}
1008: \plotone{f7e.eps}
1009: \caption{Linearly polarized and total intensity - maps of maser spots in W75N.
1010: a) Spot A: vector - linear polarization intensity (length) and relative position angle (direction);
1011: contours -- total intensity: 0.1, 0.5, 0.9 $\times$ (peak intensity).
1012: Peak values $\sqrt{Q^2+U^2}=1.6$ Jy/beam, I=16.9 Jy/beam; 								 
1013: b) Spot E: $\sqrt{Q^2+U^2}=5.4$ Jy/beam, I=5.8 Jy/beam;
1014: c) Spot F: $\sqrt{Q^2+U^2}=9.9$ Jy/beam, I=10.2 Jy/beam;
1015: d) Spots G and H: (G) $\sqrt{Q^2+U^2}=3.0$ Jy/beam, I=3.3 Jy/beam;
1016: 		(H) $\sqrt{Q^2+U^2}=2.0$ Jy/beam, I=2.9 Jy/beam;
1017: e) Spot L: $\sqrt{Q^2+U^2}=8.9$ Jy/beam, I=21.3 Jy/beam. }
1018: \end{figure}
1019: 
1020: \clearpage
1021: 
1022: \begin{figure}
1023: \epsscale{0.80}
1024: \plotone{f8.eps}
1025: \caption{ Percentage of circular polarization m$_C$ of maser spots in W75N, plotted
1026: against percentage of linear polarization m$_L$. A completely polarized feature
1027: would lie on the dotted line. Negative values of m$_C$ indicate LH circular or elliptical
1028: polarization and correspond to negative V in Table~1.   } 
1029: \end{figure} 
1030: 
1031: \clearpage
1032: 
1033: \begin{figure}
1034: \epsscale{0.80}
1035: \plotone{f9.eps}
1036: \caption{ Line profile of 1667~MHz Spot L. Solid line (filled circles): total
1037: intensity I; dashed line (triangles): percentage of linear polarization
1038: m$_L$; dashed line (squares): percentage of circular polarization m$_C$.  } 
1039: \end{figure} 
1040: 
1041: 
1042: \end{document}
1043: 
1044: 
1045: 
1046: