0704.3204/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: \usepackage{emulateapj5}
3: \begin {document}
4: 
5: \title{The Magnetic Field Structure of the LMC 2 Supershell: NGC 2100}
6: 
7: \author{John P. Wisniewski\altaffilmark{1,2,3}, Karen S. Bjorkman\altaffilmark{3,4}, 
8: Antonio M. Magalh\~aes\altaffilmark{3,5}, Antonio Pereyra\altaffilmark{5}}
9: 
10: 
11: \altaffiltext{1}{NASA GSFC Code 667, Greenbelt, MD 20771 USA, jwisnie@milkyway.gsfc.nasa.gov}
12: 
13: \altaffiltext{2}{NPP Fellow}
14: 
15: \altaffiltext{3}{Visiting Astronomer, Cerro Tololo Inter-American Observatory}
16: 
17: \altaffiltext{4}{Ritter Observatory, Department of Physics and Astronomy MS 113, University of Toledo, Toledo, OH 43606, karen.bjorkman@utoledo.edu}
18: 
19: \altaffiltext{5}{IAG, Universidade de S\~ao Paulo, Caixa Postal 3386, S\~ao Paulo, SP 01060-970, Brazil, mario@astro.iag.usp.br, antonio@astro.iag.usp.br}
20: 
21: \begin{abstract}
22: 
23: We present U,B,V,R,I imaging polarimetry of NGC 2100 and its surrounding
24: environment, which comprise a part of the LMC 2 supershell.  The morphology 
25: of the observed position angle distribution provides a tracer
26: of the projected magnetic field in this environment.  Our polarization
27: maps detail regions exhibiting similarly aligned polarization position angles,
28: as well as more complex position angle patterns.  We observe regions of
29: coherent fields on spatial scales of 42 x 24 pc to 104 x 83 pc, and infer
30: projected field strengths of $\sim$14-30$\mu$G.  We propose that the superposition of global 
31: outflows from the LMC 2 environment, as well as outflows created within NGC 2100, 
32: produce the unique field geometry in the region.  
33: 
34: \end{abstract}
35: 
36: \keywords{
37: ISM: bubbles --- ISM: magnetic fields --- Magellanic Clouds --- 
38: open clusters and associations: individual (NGC 2100) --- 
39: techniques: polarimetric --- stars: individual (HIP 21556)}
40: 
41: \section{Introduction}
42: 
43: Magnetic fields are known to align dust grains, hence influence astrophysical processes, in 
44: a wide variety of environments (see e.g. \citealt{goo96,laz03}).  The alignment of grains 
45: in the diffuse interstellar medium of 
46: the Milky Way \citep{ma70a} and the Magellanic Clouds \citep{ma70b} has long been known 
47: from linear polarization studies, which measure the dichroic absorption of starlight.  Similarly, 
48: polarimetric observations indicate grains may also be aligned in environments such 
49: as dark clouds \citep{laz97,hil99}, and the 
50: circumstellar environments of young \citep{ait93,tam99} and old \citep{ait95} stars.
51: As reviewed within \citet{laz03}, a number of alignment mechanisms have been proposed 
52: over the past sixty years, and it is likely that the local conditions of each 
53: astrophysical environment will dictate the relative importance of each of these 
54: proposed alignment mechanisms.  Characterizing the typical polarimetric 
55: behavior in each of these environments offers one avenue to constrain the relative 
56: importance of these mechanisms in each environment.   
57: 
58: Massive stars in OB associations provide a rich source of
59: outflows from stellar winds and supernovae explosions; the interaction of
60: such outflows with the surrounding interstellar medium (ISM) is known to
61: create large supershells.  Observations suggest that some of these shells 
62: may be magnetized, with typical fields strengths of at least several to tens of $\mu$G 
63: \citep{val93,val94,per07}, and it is expected that these fields can influence the 
64: dynamical evolution of shells, including constraining their expansion 
65: \citep{min93,tom98}.  With a diameter of $\sim$ 900 pc,
66: LMC 2 was initially identified as a supershell based upon its
67: morphology \citep{mea80}.  The
68: kinematics of the LMC 2 supershell has been a subject of some debate: 
69: \citet{cau82} suggested that LMC 2 was expanding as a cohesive structure; 
70: however, followup studies by \citet{mea87}, \citet{poi99} and \citet{amb04} 
71: argue against a 
72: global expansion, asserting the structure is a conglomeration of localized
73: expanding structures.  Observed extended X-ray emission in the region led
74: \citet{wan91} to suggest that LMC 2 was formed as a result of a
75: superbubble breaking out from the plane of the Large Magellanic Cloud (LMC).  
76: 
77: The richest of the 29 stellar associations in the direction of LMC 2 
78: identified by \citet{hod67} projected to lie interior
79: to LMC 2 is NGC 2100. Given its age, $\sim$15 Myr \citep{cas96}, 
80: \citet{poi99} has postulated that all of NGC 2100's O stars have exploded as
81: supernovae, implying that it has played an important role in shaping the
82: observed structure of LMC 2.  \citet{poi99} and
83: \citet{amb04} observed complex H$\alpha$ velocity components which vary across
84: the extent of NGC 2100, suggesting the presence of a turbulent interstellar
85: environment.  The interstellar polarization near NGC 2100 is also known to differ 
86: from general behavior of the LMC (0.32 - 0.57\% at position angles of 28-45$^{\circ}$, 
87: \citealt{wi05a,wi05b}); \citet{ma70b} observed a strong focusing of magnetic lines 
88: around 30 Doradus (30 Dor), which is located $\sim$16$\farcm$ west of the LMC 2 supershell.  
89: 
90: In this paper, we present detailed polarization maps of NGC 2100 and its surrounding 
91: environment, which reveal evidence of a complex magnetic field morphology.  These 
92: data provide diagnostics of the grain alignment mechanisms which likely dominate in 
93: such dynamic astrophysical environments.  In Section 2, we
94: describe our polarimetric observations.  Polarization maps and estimates
95: of projected magnetic field strengths are presented in Section 3.  In
96: Section 4, we present a discussion of these results and the clues they
97: may offer towards understanding the formation and evolution of the 
98: LMC 2 supershell.
99: 
100: 
101: \section{Observations}
102: 
103: Imaging polarimetry of NGC 2100 and its surrounding field were
104: obtained at the CTIO 1.5 m telescope, as summarized in Table 1.  We used
105: the F/7.5 secondary configuration, yielding a 15$\farcm$0 field of view and
106: a 0$\farcs$44 pixel$^{-1}$ scale.  Data were recorded with the telescope's
107: standard Cassegrain focus CCD (CFCCD), a 2048 x 2048 CCD which was
108: read out in dual amplifier mode.  The standard telescope configuration was
109: modified by the addition of a rotatable half-wave plate, followed by a
110: dual calcite block (Savart plate) placed in the first filter wheel.  This dual calcite 
111: analyzer, whose optical axes were crossed to minimize astigmatism and color
112: effects, simultaneously produced two orthogonally polarized images of 
113: all objects, allowing for the near complete cancellation of all background
114: sky polarization, as well as atmospheric transparency effects \citep{mag96}.  
115: A small imperfection in the calcite
116: block was masked out, shrinking the effective field of view of the instrument
117: by 2$\farcm$0-3$\farcm$0 arcminutes in the southeast corner of the CCD and 
118: by $\sim$1$\farcm$5 in the 
119: northeast and northwest corners of the chip.  Standard Johnson U, B, V, R, 
120: and I filters
121: were housed in the second filter wheel.  Images were taken at 8 wave-plate
122: positions, each separated by 22.5$^{\circ}$, allowing us to derive full
123: linear polarization measurements for NGC 2100.  Additional information
124: regarding this instrument can be found in \citet{mag96}, \citet{mel01}, 
125: and \citet{per02}.
126: 
127: Basic image processing was done in IRAF\footnote{IRAF
128: is distributed by the National Optical Astronomy Observatories, which are
129: operated by the Association of Universities for Research in Astronomy, Inc.,
130: under contract with the National Science Foundation.} using standard 
131: techniques.  After deriving aperture photometry for our images, the least 
132: squares solution of the 8 wave-plate positions, calculated with the
133: PCCDPACK polarimetric reduction package \citep{per00}, yielded linear
134: polarization measurements.  The residuals at each wave-plate position, $\psi_{i}$, with 
135: respect to the expected $\cos 4\psi_{i}$ curve constitute the uncertainties in 
136: our data; these are consistent with the theoretically expected photon noise
137: errors \citep{mag84}.
138: 
139: Instrumental polarization effects were determined from observations
140: of polarized and unpolarized standard stars, obtained nightly during our
141: ten day observing run in 2001.  These data are self-consistent and
142: agree with observations obtained in a subsequent eleven day observing
143: run in 2002 \citep{wis03} using the same instrument, illustrating 
144: its excellent stability.  The instrumental polarization was measured to be within 
145: 0.03\% (I filter) to 0.07\% (B filter), thus no correction was applied to
146: our data.  Note that due to a lack of known faint unpolarized standard stars,  
147: we used the M star HIP 21556 as an unpolarized standard star for the 
148: B, V, R, and I filters.  Given its nearby location (d=11 pc) and spectral type,
149: one would not expect such an object to exhibit significant 
150: polarization \citep{tin82}.
151: Indeed, observations of this target during our 2001 and 2003 observing runs
152: showed it to be unpolarized. 
153: 
154: \section{Results}
155: 
156: The polarization data we discuss represent the superposition of two components
157: of distinctly different origin, namely interstellar and intrinsic polarization.
158: Intrinsic polarization can arise from a variety of scattering mechanisms
159: within the circumstellar environment of a host star, while interstellar 
160: polarization results from dichroic absorption of starlight by aligned
161: interstellar dust grains located along the line of sight.   
162: Most of our targets should be normal main-sequence stars which are not characterized 
163: by the presence of an extended (or asymmetrical) circumstellar envelope, hence 
164: they will exhibit no intrinsic polarization at the precision level
165: of our measurements.  The statistical analysis of the total polarization observed 
166: for each of our targets should thus provide an accurate diagnostic of the interstellar 
167: polarization along the line of sight \citep{mcl79,per02,wis03,wi05b}. 
168: Furthermore, given the LMC's large distance of 50 kpc \citep{fea91}, any
169: spatial variability we detect in this interstellar polarization component must  
170: be the result of a change in the magnetic field or interstellar dust grain
171: properties within the LMC, rather than a projection of Galactic interstellar
172: medium properties.
173: 
174: \subsection{Polarization Maps}
175: 
176: In Figures 1-5 we present the polarization in the U, B, V, R, and I filters
177: for NGC 2100 and its surrounding field.  Polarization vectors are overplotted
178: on Digitized Sky Survey2 (DSS2) red (V, R, and I filters) and blue (U and B
179: filters) images which span 0.5 square degrees, allowing one to place the
180: polarization of NGC 2100 in the context of the LMC 2 supershell.  A more
181: detailed image of LMC 2's nebulosity, including identification of the major OB 
182: associations in the area, can be seen in Figure 1c of \citet{poi99}.  
183: To exclude likely spurious detections, we have only plotted objects
184: with the following properties: $ 0.1\% <$ polarization $< 3.0\%$ 
185: and $p/\sigma_{p} > 3.0$.  
186: 
187: Numerous trends in the morphology of the polarization vectors in
188: each of the filters can immediately be seen.  The magnitude of a typical
189: polarization vector is $\sim 1.5\%$, which is significantly higher than the
190: average polarization observed throughout the LMC \citep{wi05a,wi05b}.  
191: While polarization position angles (PA) tend to be coherent
192: on small spatial scales, large-scale variability across the field of view
193: of the data set is immediately apparent.  The patterns traced by this 
194: large-scale variability are consistent across every filter, indicating the
195: phenomena represent real features.  
196: 
197: To further explore these large-scale position angle trends, we divided
198: our field of view into 5 smaller spatial scales which each seemed to possess
199: one unique, average position angle.  We assign the arbitrary labels A-E
200: to these fields, and show these fields in the R filter in Figures 6-10.  Note that we have 
201: only plotted all objects with $ 0.1\% <$ polarization $< 3.0\%$
202: and $p/\sigma_{p} > 3.0$ in these figures.
203: The mean position angle, FHWM of gaussian
204: fits to the samples, standard deviation (in radians) of objects within the gaussian
205: fit for each area, and spatial extent of each area, assuming a distance of
206: 50 kpc, are tabulated in Table 2.  The polarization position angle histograms used to 
207: derive these parameters are also given in Figures 6-10.
208: 
209: The position angle rotation to the north, west, and south of NGC 2100, 
210: apparent by casual inspection of Figures 1-5, 
211: is indeed real as we measure the mean PA to vary from 
212: 76$^{\circ}$ to 122$^{\circ}$ to 94$^{\circ}$ in Figures 7-9 respectively.
213: Figure 10 depicts spatial area E, immediately west of NGC 2100, showing
214: PA alignment at 167$^{\circ}$.  The dramatic curvature in the field pattern
215: traced out in areas B-E in the area west of NGC 2100 
216: qualitatively matches a similar ``bubble-like'' pattern seen in the 
217: H$\alpha$ image of the LMC 2 region of \citet{poi99}, e.g. their Figure 1a.
218: 
219: Systematic alignment in the eastern portion
220: of our field of view is less dramatic, as seen in Figure 6 which depicts
221: area A.  We find position angles in area A tend to be $< 90^{\circ}$, 
222: and find suggestive evidence that the PA distribution in this spatial region
223: may be fit by two gaussians at 29$^{\circ}$ (FWHM = 25$^{\circ}$) and
224: 64$^{\circ}$ (FWHM = 22$^{\circ}$) respectively.  The constituents of these
225: two possible gaussian
226: distributions occupy no unique spatial regions: it is possible that the
227: observed distributions originate at slightly different distances within the
228: LMC 2 neighborhood, projecting themselves onto common spatial regions.
229: Alternatively, it is possible that large-scale LMC-2 flows to the west of NGC 2100 are 
230: interacting with velocity fields from within NGC 2100 to produce the observed apparent 
231: superposition effects. 
232: 
233: The projected spatial extent of regions with definitive, coherent position 
234: angle alignment varies from
235: 42 x 24 pc, associated with region E, to 104 x 83 pc, associated with region
236: B.  We note the suggestive presence of smaller alignment trends within
237: and outside of some of these designated areas.  Due to small number 
238: statistics, it is unclear whether such features are real, hence illustrating
239: common alignment on finer scales, or whether they are the
240:  result of small intrinsic
241: polarization components adding small scatter to the data set.  Deeper
242: polarimetric mapping of the region would clarify the presence of small-scale
243: alignment by providing a larger statistical database.  If our polarization
244: maps are indeed tracing some of the larger structures in the H$\alpha$ maps
245: of \citet{poi99}, then we would expect deeper polarization maps to trace
246: many of the finer nebulosity seen in such images.
247: 
248: 
249: \subsection{Estimating B Fields}
250: 
251: A formalism for estimating magnetic field strengths from polarimetric
252: observations was developed by \citet{cha53}(C-F technique), and has been since modified
253: to account for various inadequacies 
254: (see e.g. \citet{goo96,zwe96,hei01,cru04}).  As summarized by \citet{hei01,hen01} and \citet{per07}, while the C-F method is a commonly used technique to estimate magnetic 
255: field strengths, its use is dubious in cases of, amongst other 
256: factors, large polarization position angle dispersions and large turbulent velocity 
257: dispersions.  We have used the description
258: provided by equation 7 of \citet{hei01}, \begin{equation} 
259: B = \frac{1}{2}\sqrt{4 \pi \rho \left( \begin{array}{c} \frac{\sigma(v_{los})^{2}}
260: {\sigma(tan \delta)^{2}} \end{array} \right)}
261: \end{equation}, where $\rho$ is the mean density, $\sigma(v_{los})$ is the
262: dispersion in the line-of-sight velocity, and $\sigma(tan \delta)$ is the dispersion in polarization position angles, i.e. the difference, within the distribution, between the 
263: position angle of a given object and the average position angle.   This 
264: description was used as
265: it eliminates the small angle approximation present
266: in the original formalism; furthermore, it includes a factor of 1/2 to account for 
267: the field overestimation provided
268: by the classical Chandrasekhar-Fermi method \citep{cru04}.
269: 
270: We were able to measure position angle dispersions for regions B-E of our
271: dataset using PCCDPACK and tabulate the standard deviation of these values, 
272: $\sigma(tan\delta)$ in Table 2.  \citet{poi99} reported a
273: HI number density of 3-4 cm$^{-3}$: we assumed a number density of 4
274: cm$^{-3}$ in our calculations.  We estimated the line of sight velocity for our 
275: regions from the HI velocities reported by \citet{mea87} for their region 43, 
276: corresponding to the approximate location of NGC 2100, $\sigma_{vlos}$ = 
277: 52 km s$^{-1}$.  This dispersion is consistent with the FWHM of H$\alpha$ 
278: velocities reported by \citet{poi99} across their E-I spectroscopic cut.  
279: The resulting
280: magnetic field strengths for our fields range from 14-30 $\mu$G, as tabulated
281: in Table 2.  We stress that these field values should only be considered
282: crude estimates: detailed measurements of the gas velocity dispersions and
283: densities corresponding to the specific spatial regions in which we observed
284: polarization position angle dispersions are needed to further refine these
285: field estimates.  Nevertheless, our derived range of field strengths are consistent with the
286: strength of random field fluctuations in the LMC reported by \citet{gae05}, especially 
287: those located nearby supernova remnants and wind bubbles, which were quoted to 
288: be $\sim$8 $\mu$G by these authors.
289: 
290: 
291: 
292: \section{Discussion}
293: 
294: We now explore some of the implications of the polarization maps presented in
295: Section 3.1.  Our polarization maps of NGC 2100 and its nearby environment 
296: indicate the
297: presence of a sizable magnitude of interstellar polarization, $\sim1.5\%$
298: which experiences systematic position angle changes.  We attribute this
299: position angle variability to changes in the orientation of the projected
300: magnetic field.  It is equally likely that the third dimension of this
301: field also varies.  The only effect such a variation would have on our data set
302: would be additional dispersion in the distribution of polarization levels.  
303: While a wide distribution is observed, 
304: other factors such as the presence of small intrinsic polarization components 
305: in objects, small variability in the distance of objects within our 
306: field of view,
307: and changes in the polarizing properties of the grains across the field of 
308: view, i.e. changes in grain size, shape, or composition, also likely serve
309: to broaden the observed distribution.  Additional observational tools, 
310: such as that provided by atomic alignment \citep{yan06}, could be used 
311: to provide an independent measure of the localized three-dimensional 
312: magnetic field.  
313: 
314: \subsection{Origin of Position Angle Variations}
315: 
316: We consider the origin of the position angle variability detailed
317: in Section 3.1.  Given the dynamic nature of the region,
318: it seems plausible to expect the complex field patterns, which align grains to
319: produce the observed polarization, to be driven by the various outflows 
320: present.  Such a scenario is supported by the possible tracing of large 
321: H$\alpha$ features by our data, as noted in Section 3.1.  The morphological
322: details of Figures 1-5 suggest that the field patterns are not solely
323: guided by the stellar outflows and supernovae remnants of NGC 2100.
324: Studies of the interstellar polarization
325: surrounding other young LMC clusters and OB associations with similar hot
326: star contents do not illustrate these types of complex field 
327: patterns \citep{wi05a,wi05b}.  Within the LMC 2 environment, the western side of
328: NGC 2100 shows complex field patterns while the eastern side of the 
329: cluster only displays moderate evidence of cohesive field alignment.  No
330: asymmetry in the distribution of massive stars or their remnants in this
331: cluster has been observed, thus we don't expect winds or outflows from 
332: NGC 2100's massive star population to be responsible for producing these
333: field patterns.  Rather, we speculate that other large-scale flows might play a major role 
334: in twisting field lines in the observed patterns.
335: 
336: \citet{poi99} suggest that rather than being a cohesively expanding shell, 
337: the geometry of LMC 2 is that of two HI sheets enclosing a region of hotter
338: gas.  They suggest both
339: cavity material and the surface of the HI sheets are being swept eastward across
340: the complex by the outflows of material located on the western edge of the
341: region.  We speculate that such general, large-scale flows, carrying
342: with them local magnetic field lines, might move past
343: the northern and southern boundaries of NGC 2100 and be impeded by the cluster
344: itself.  Such a scenario could
345: account for the coherent position angle patterns located to the north and south
346: of NGC 2100, as well as the dramatic turnabout in the field immediately west
347: of the cluster.  As the neighborhood due east of NGC 2100 would be partially
348: shielded from such flows, one would expect less coherent field patterns
349: in this environment, as is observed to the east of NGC 2100.
350: The 30 Dor complex, a rich site of powerful stellar outflows, is
351: located to the west of NGC 2100.  From polarization measurements, 
352: \citet{ma70b} noted strong magnetic 
353: focusing in the 30 Dor region.  Thus we suggest 30 Dor should be considered
354: as a possible source of outflows which influence LMC 2 and shape the 
355: projected magnetic field patterns we observe.
356: 
357: A number of grain alignment mechanisms have been postulated and, as summarized in 
358: the review paper of \citet{laz03}, it is likely that different mechanisms may dominate in different astrophysical environments, depending upon the local conditions present.  
359: Some of the proposed mechanisms include 
360: the Davis-Greenstein process, in which paramagnetic dissipation by rotating 
361: grains leads to alignment 
362: \citep{dav51}, the Gold process, in which grains are mechanically aligned via 
363: collisional interactions with a supersonic gas flow \citep{gol52,laz94,la97b}, 
364: and radiative torques, in which alignment is achieved via the spin-up of irregularly shaped grains which scatter left- and right-hand polarized light in a different way \citep{dol76,dra96,dra97}.  The LMC-2 supershell itself, if it assumed to be a cohesively 
365: expanding body (e.g. \citealt{cau82}), is only characterized by an expansion velocity of $\sim$30 km s$^{-1}$ \citep{cau82}, which is well below the supersonic gas velocity 
366: required for the Gold mechanical alignment mechanism.  However,  
367: NGC 2100's proximity to both the 30 Dor region and the winds of massive stars within 
368: NGC 2100 suggest that local grains might interact with a more dynamic gas flow than 
369: that which characterizes the much larger LMC-2 region.  As such, we suggest that 
370: mechanical alignment might indeed play a partial role in constructing the observed 
371: morphology of aligned grains in the NGC 2100 region; clearly detailed modeling of 
372: the system would be advantageous to quantitatively constrain the various grain alignment 
373: mechanisms which could be operating in this dynamic environment. 
374: 
375: 
376: \subsection{Location of Polarizing Region}
377: 
378: While we have interpreted the bulk of the magnetic field variability implied
379: by our observations to be tied to the dynamics of the inner layer of LMC 2, 
380: we now consider the possible influence of the HI sheets which encompass 
381: this layer in
382: the proposed 3-dimensional picture of \citet{poi99}.  Based upon the
383: derived total line of sight reddening for NGC 2100, $E_{B-V}$ = 0.24 \citep{kel00}, the standard
384: relationship between polarization and extinction presented in \citet{ser75},
385: $P_{max} <$ 9 E$_{B-V}$, predicts an interstellar polarization of $< 2.2\%$, in
386: agreement with the average magnitude observed in our data set, $\sim1.5\%$.
387: \citet{poi99} report the thin (80-100 pc) HI sheets have 
388: column densities of $\sim$1 x 10$^{21}$ cm$^{-2}$.  This column density implies
389: a reddening value similar to that of \citet{kel00}, hence a similar predicted
390: maximum magnitude of interstellar polarization, based upon the relation
391: $N_{HI} / E_{B-V}$ = 5 x 10$^{21}$ atoms cm$^{-2}$ mag$^{-1}$ \citep{sav72}.
392: Thus it appears that enough dichroic absorption by interstellar dust grains
393: could occur within the thin HI sheet positioned in front of NGC 2100 to
394: produce the level of observed polarization.  
395: 
396: 
397: \subsection{Summary}
398: 
399: We have presented polarization maps for a subsection of the LMC 2 supershell,
400: namely NGC 2100 and its surround field.  These maps show regions of aligned
401: position angles on scales of 42 x 24 pc to 104 x 83 pc, attributable
402: to absorption by interstellar dust grains aligned by projected magnetic
403: fields.  We estimate these projected fields to have strengths of 8-17
404: $\mu$G, and stress that more accurate field estimates may be achieved by
405: incorporating measurements which better reflect the interstellar medium
406: properties corresponding to our survey area.  A plausible explanation for the
407: observed complex field patterns is that outflows present within LMC 2, modified by 
408: velocity fields from NGC 2100, combine to produce the observed field patterns. 
409: The observed asymmetrical field morphology suggests the
410: stellar sources in NGC 2100 are not the primary source of outflows shaping
411: the observed fields.  Rather, we speculate that NGC 2100 may serve to 
412: disrupt the path of large-scale flows moving eastward across LMC 2.  
413: We suggest that the 30 Dor region, observed to be a source
414: of both massive outflows and strong magnetic fields, may be the source 
415: powering the observed field patterns in LMC 2.
416: 
417: Finally, we considered a proposed 3-dimensional picture of LMC 2 in which two  
418: HI shells confine a region of hotter gas.  We find the magnitude of observed
419: polarization could be produced by aligned dust grains within
420: one of these HI shells, noting that some mechanism must then impart the
421: complex field geometry produced within the inner gas layer to this thin outer
422: shell.
423: 
424: 
425: \acknowledgments
426: 
427: We thank the anonymous referee whose comments helped to improve this 
428: paper.  This research was supported by NASA NPP and GSRP fellowships to JPW  
429: (NNH06CC03B, NGT5-50469), a NASA LTSA grant (NAG5-8054) and a 
430: Research Corporation Cottrell Scholar award to KSB, and a FAPESP grant 
431: (02/12880-0) to AP.  AMM also acknowledges support from the Brazillian 
432: agencies FAPESP and CNPq.  Polarimetry at the University of S\~ao Paulo (USP) 
433: is supported by FAPESP.  This  
434: research has made use of NASA's Skyview virtual observatory, NASA ADS, and
435: the SIMBAD database.
436: 
437: 
438: \begin{thebibliography}{}
439: 
440: \bibitem[Aitken et al.(1993)]{ait93} Aitken, D.K., Wright, C.M., Smith, C.H., \& Roche, P.F. 
441: 1993, MNRAS, 262, 456
442: 
443: \bibitem[Aitken et al.(1995)]{ait95} Aitken, D.K., Smith, C.H., Moore, T.J.T., \& Roche, P.F. 
444: 1995, MNRAS, 273, 359
445: 
446: \bibitem[Ambrocio-Cruz et al.(2004)]{amb04} Ambrocio-Cruz, P. et al. 2004, 
447: AJ, 127, 2145
448: 
449: \bibitem[Cassatella et al.(1996)]{cas96} Cassatella, A., Barbero, J., 
450: Brocato, E., Catellani, V., \& Geyer, E.H. 1996, A\&A, 306, 125
451: 
452: \bibitem[Caulet et al.(1982)]{cau82} Caulet, A., Deharveng, L., Georgelin, Y.
453: P., \& Georgelin, Y.M. 1982, A\&A, 110, 185
454: 
455: \bibitem[Chandrasekhar \& Fermi(1953)]{cha53} Chandrasekhar, S. \& Fermi, E. 
456: 1953, ApJ, 118, 113
457: 
458: \bibitem[Crutcher(2004)]{cru04} Crutcher, R.M. 2004, The Magnetized 
459: Interstellar Medium, ed. B. Uyaniker, W. Reich, \& R. Wielebinski, 123
460: 
461: \bibitem[Davis \& Greenstein(1951)]{dav51} Davis, L. \& Greenstein, J.L. 1951, ApJ, 114, 206
462: 
463: \bibitem[Dolginov \& Mytrophanov(1976)]{dol76} Dolginov, A.Z. \& Mytrophanov, I.G. 1976, 
464: Ap\&SS, 43, 291
465: 
466: \bibitem[Draine \& Weingartner(1996)]{dra96} Draine, B.T. \& Weingartner, J.C. 1996, ApJ, 470, 551
467: 
468: \bibitem[Draine \& weingartner(1997)]{dra97} Draine, B.T. \& Weingartner, J.C. 1997, ApJ, 
469: 480, 633
470: 
471: \bibitem[Feast(1991)]{fea91} Feast, M.W. 1991, in IAU Sump. 148, The
472: Magellanic Clouds, ed. R. Haynes, \& D. Milne (Dordrecht: Kluwer), 1
473: 
474: \bibitem[Gaensler et al.(2005)]{gae05} Gaensler, B.M., Haverkorn, M., Staveley-Smith, L., 
475: Dickey, J.M., McClure-Griffiths, N.M., Dickel, J.R., \& Wolleben, M. 2005, Science, 
476: 307, 1610
477: 
478: \bibitem[Gold(1952)]{gol52} Gold, T. 1952, MNRAS, 112, 215
479: 
480: \bibitem[Goodman(1996)]{goo96} Goodman, A.A. 1996, in ASP Conf. Ser. 97, 
481: Polarimetry of the Interstellar Medium, ed. W.G. Roberge \& D.C.B. Whittet 
482: (San Francisco: ASP), 325
483: 
484: \bibitem[Heitsch et al.(2001)]{hei01} Heitsch, F., Zweibel, E.G., Mac Low, 
485: M.-M., Li, P., \& Norman, M.L. 2001, ApJ, 561, 800
486: 
487: \bibitem[Henning et al.(2001)]{hen01} Henning, Th., Wolf, S., Launhardt, R., \& 
488: Waters, R. 2001, ApJ, 561, 871
489: 
490: \bibitem[Hildebrand et al.(1999)]{hil99} Hildebrand, R.H., Dotson, J.L., Dowell, C.D., 
491: Schleuning, D.A., \& Vaillancourt, J.E. 1999, ApJ, 516, 834
492: 
493: \bibitem[Hodge \& Wright(1967)]{hod67} Hodge, P.W. \& Wright, F.W. 1967, 
494: Smithsonian Pub. 4699, The Large Magellanic Cloud (Washington: Smithsonian
495: Press)
496: 
497: \bibitem[Keller et al.(2000)]{kel00} Keller, S.C., Bessell, M.S., \& Da Costa,
498: G.S. 2000, AJ, 119, 1748
499: 
500: \bibitem[Lazarian(1994)]{laz94} Lazarian, A. 1994, MNRAS, 268, 713
501: 
502: \bibitem[Lazarian et al.(1997)]{laz97} Lazarian, A., Goodman, A.A., \& Myers, P.C. 
503: 1997, ApJ, 490, 273
504: 
505: \bibitem[Lazarian(1997)]{la97b} Lazarian, A. 1997, ApJ, 483, 296
506: 
507: \bibitem[Lazarian(2003)]{laz03} Lazarian, A. 2003, JQSRT, 79, 881
508: 
509: \bibitem[Magalh\~aes, Benedetti, \& Roland(1984)]{mag84} Magalh\~aes, A.M., Benedetti, E., \& Roland, E. 1984, PASP, 96, 384
510: 
511: \bibitem[Magalh\~aes et al.(1996)]{mag96} Magalh\~aes, A.M, Rodrigues, C.V., 
512: Margoniner, V.E., \& Pereyra, A. 1996, in ASP Conf. Ser. 97, Polarimetry of the
513: Interstellar Medium, ed. W.G. Roberge \& D.C.B. Whittet (San Francisco: ASP), 
514: 118
515: 
516: \bibitem[Mathewson \& Ford(1970a)]{ma70a} Mathewson, D.S. \& Ford, V.L. 1970a,
517: MmRAS, 74, 139
518: 
519: \bibitem[Mathewson \& Ford(1970b)]{ma70b} Mathewson, D.S. \& Ford, V.L. 1970b, 
520: ApJ, 160L, 43
521: 
522: \bibitem[McLean \& Clarke(1979)]{mcl79} McLean, I.S. \& Clarke, D. 1979, MNRAS, 
523: 186, 245
524: 
525: \bibitem[Meaburn(1980)]{mea80} Meaburn, J. 1980, MNRAS, 192, 365
526: 
527: \bibitem[Meaburn et al.(1987)]{mea87} Meaburn, J., Marston, A.P., McGee, R.X.,
528:  \& Newton, L.M. 1987, MNRAS, 225, 591
529: 
530: \bibitem[Melgarejo et al.(2001)]{mel01} Melgarejo, R., Magalh\~aes, A.M., Carciofi, A.C., \& Rodrigues, C.V. 2001, A\&A, 377, 581
531: 
532: \bibitem[Mineshige et al.(1993)]{min93} Mineshige, S., Shibata, K., \& Shapiro, P.R. 1993, 
533: ApJ, 409, 663
534: 
535: \bibitem[Pereyra(2000)]{per00} Pereyra, A. 2000, Ph.D. thesis, Univ. S\~ao 
536: Paulo
537: 
538: \bibitem[Pereyra \& Magalh\~aes(2002)]{per02} Pereyra, A. \& Magalh\~aes, 
539: A.M. 2002, ApJS, 141, 469
540: 
541: \bibitem[Pereyra \& Magalh\~aes(2007)]{per07} Pereyra, A. \& Magalh\~aes, 
542: A.M. 2007, ApJ, in press (astro-ph/0702550)
543: 
544: \bibitem[Points et al.(1999)]{poi99} Points, S.D. et al. 1999, ApJ, 518, 298
545: 
546: \bibitem[Savage \& Jenkins(1972)]{sav72} Savage, B.D. \& Jenkins, E.B. 1972, 
547: ApJ, 172, 491
548: 
549: \bibitem[Serkowski et al.(1975)]{ser75} Serkowski, K., Mathewson, D.S.,
550: \& Ford, V.L. 1975, ApJ, 196, 261
551: 
552: \bibitem[Tamura et al.(1999)]{tam99} Tamura, M., Hough, J.H., Greaves, J.S., Morino, J-I.,
553: Chrysostomou, A., Holland, W.S., \& Momose, M. 1999, ApJ, 525, 832
554: 
555: \bibitem[Tinbergen(1982)]{tin82} Tinbergen, J. 1982, A\&A, 105, 35
556: 
557: \bibitem[Tomisaka(1998)]{tom98} Tomisaka, K. 1998, MNRAS, 298, 797
558: 
559: \bibitem[Vallee(1993)]{val93} Vallee, J.P. 1993, ApJ, 419, 670
560: 
561: \bibitem[Vallee(1994)]{val94} Vallee, J.P. 1994, Ap\&SS, 220, 289
562: 
563: \bibitem[Wang \& Helfand(1991)]{wan91} Wang, Q. \& Helfand, D.J. 1991, ApJ, 
564: 379, 327
565: 
566: \bibitem[Wisniewski et al.(2003)]{wis03} Wisniewski, J.P., Bjorkman, K.S., 
567: \& Magalh\~aes, A.M. 2003, ApJL, 598, 43
568: 
569: \bibitem[Wisniewski(2005)]{wi05a} Wisniewski, J.P. 2005, Ph.D. thesis,
570: University of Toledo
571: 
572: \bibitem[Wisniewski et al.(2007)]{wi05b} Wisniewski, J.P., Bjorkman, K.S., 
573: Magalh\~aes, A.M., Meade, M.R., \& Pereyra, A. 2007, ApJ, submitted
574: 
575: \bibitem[Yan \& Lazarian(2006)]{yan06} Yan, H. \& Lazarian, A. 2006, ApJ, 653, 1292
576: 
577: \bibitem[Zweibel(1996)]{zwe96} Zweibel, E.G. 1996, in ASP Conf. Ser. 97, 
578: Polarimetry of the Interstellar Medium, ed. W.G. Roberge \& D.C.B. Whittet
579: (San Francisco: ASP), 486
580: 
581: 
582: \end{thebibliography}
583: 
584: \clearpage
585: %\newpage
586: \begin{table}
587: \caption{Journal of NGC 2100 Observations}
588: \begin{tabular}{lcc}
589: 
590: Filter & Obs. Date & Exposure Time \\
591: 
592: \tableline
593: 
594: U & 24 Nov. 2001 & 1200 sec. \\
595: B &  23 Nov. 2001 & 240 sec. \\
596: V & 23 Nov. 2001 & 180 sec. \\
597: R & 24 Nov. 2001 & 180 sec. \\
598: I & 23 Nov. 2001 & 180 sec. \\
599: 
600: \tablecomments{Note that the listed exposure times 
601: correspond to the total integration at each of 8 wave-plate positions.}
602: 
603: \end{tabular}
604: \end{table}
605: 
606: 
607: \clearpage
608: %\newpage
609: \begin{table}
610: \caption{Summary of Magnetic Field Properties by Region}
611: \begin{tabular}{lcccccc}
612: 
613: Region & \# Stars & Spatial Extent (pc) & Mean PA (deg) & FWHM (deg) & $\sigma(tan\delta)$ (rad) & B ($\mu$G) \\
614: 
615: \tableline
616: 
617: A & 105 & 73 x 177 & 0-90 & \nodata & \nodata & \nodata \\
618: B & 110 & 104 x 83 & 76 & 24 & 0.24 & 25 \\
619: C & 88 & 63 x 75  & 122 & 37 & 0.26 & 23 \\
620: D & 56 & 104 x 63 & 94 & 22 & 0.20 & 30 \\
621: E & 56 & 42 x 24  & 167 & 46 & 0.44 & 14 \\
622: 
623: \tablecomments{Summary of the polarization position angle variability
624: across our field of view.}
625: 
626: 
627: \end{tabular}
628: \end{table}
629: 
630: 
631: \clearpage
632: 
633: \begin{figure}
634: \includegraphics[scale=0.5]{newf1.eps}
635: \caption{U filter polarization of 462 targets which have 
636: p$/\sigma_{p} > 3$ in the vicinity of NGC 2100.  Polarization vectors
637: are overplotted on a 0.5 deg$^{2}$ DSS2 blue image.}
638: \end{figure}
639: 
640: \begin{figure}
641: \includegraphics[scale=0.5]{newf2.eps}
642: \caption{B filter polarization of 592 targets which have 
643: p$/\sigma_{p} > 3$ in the vicinity of NGC 2100.  Polarization vectors
644: are overplotted on a 0.5 deg$^{2}$ DSS2 blue image.}
645: \end{figure}
646: 
647: \begin{figure}
648: \includegraphics[scale=0.5]{newf3.eps}
649: \caption{V filter polarization of 661 targets which have 
650: p$/\sigma_{p} > 3$ in the vicinity of NGC 2100.  Polarization vectors
651: are overplotted on a 0.5 deg$^{2}$ DSS2 red image.}
652: \end{figure}
653: 
654: \begin{figure}
655: \includegraphics[scale=0.5]{newf4.eps}
656: \caption{R filter polarization of 700 targets which have 
657: p$/\sigma_{p} > 3$ in the vicinity of NGC 2100.  Polarization vectors
658: are overplotted on a 0.5 deg$^{2}$ DSS2 red image.}
659: \end{figure}
660: 
661: \begin{figure}
662: \includegraphics[scale=0.5]{newf5.eps}
663: \caption{I filter polarization of 621 targets which have 
664: p$/\sigma_{p} > 3$ in the vicinity of NGC 2100.  Polarization vectors
665: are overplotted on a 0.5 deg$^{2}$ DSS2 red image.}
666: \end{figure}
667: 
668: \begin{figure}
669: \includegraphics[scale=0.4]{newf6a.eps}
670: \includegraphics[scale=1.0]{f6b.eps}
671: \caption{Section A of our field of view, showing the polarization of 105
672: objects with 0.1\% $<$ p $<$ 3.0\% and p$/\sigma_{p} > 3$.  We find the
673: position angles in this region are concentrated at angles $<$ 90$^\circ$,
674: possibly following a bimodal distribution with centers at 29$^\circ$ and
675: 64$^\circ$.}
676: \end{figure}
677: 
678: \begin{figure}
679: \includegraphics[scale=0.4]{newf7a.eps}
680: \includegraphics[scale=0.29]{f7b.eps}
681: \caption{Section B of our field of view, showing the polarization of 110
682: objects with 0.1\% $<$ p $<$ 3.0\% and p$/\sigma_{p} > 3$.  The mean 
683: polarization position angle of these stars, determined by a gaussian fit,
684: is 76$^{\circ}$.} 
685: \end{figure}
686: 
687: \begin{figure}
688: \includegraphics[scale=0.4]{newf8a.eps}
689: \includegraphics[scale=0.7]{f8b.eps}
690: \caption{Section C of our field of view, showing the polarization of 88
691: objects with 0.1\% $<$ p $<$ 3.0\% and p$/\sigma_{p} > 3$.  The mean 
692: polarization position angle of this region was determined to be 122$^{\circ}$.}
693: \end{figure}
694: 
695: \begin{figure}
696: \includegraphics[scale=0.4]{newf9a.eps}
697: \includegraphics[scale=1.0]{f9b.eps}
698: \caption{Section D of our field of view, showing the polarization of 56
699: objects with 0.1\% $<$ p $<$ 3.0\% and p$/\sigma_{p} > 3$.  The mean 
700: polarization position angle of this region was determined to be 94$^{\circ}$.}
701: \end{figure}
702: 
703: \begin{figure}
704: \includegraphics[scale=0.4]{newf10a.eps}
705: \includegraphics[scale=1.0]{f10b.eps}
706: \caption{Section E of our field of view, showing the polarization of 56
707: objects with 0.1\% $<$ p $<$ 3.0\% and p$/\sigma_{p} > 3$.  The mean
708: polarization position angle of this region was determined to be 167$^{\circ}$.}
709: \end{figure}
710: 
711: 
712: 
713: 
714: 
715: \end{document}
716: