0811.0512/ms.tex
1: \documentclass[preprint]{aastex}
2: 
3: %\usepackage{aaspsfonts,graphicx}
4: \usepackage{graphicx}
5: 
6: \DeclareMathAlphabet{\mathbf}{OT1}{cmr}{bx}{it}
7: 
8: \arraycolsep=1.5pt
9: 
10: \begin{document}
11: 
12: \title{Polarimetric Diagnostics of Unresolved Chromospheric Magnetic 
13: Fields}
14: 
15: \author{R. Casini,$^a$ R. Manso Sainz,$^b$ B. C. Low$^a$}
16: 
17: \affil{$^a\,$High Altitude Observatory,
18: National Center for Atmospheric Research,\altaffilmark{1}
19: P.O.~Box 3000, Boulder, CO 80307-3000\\
20: $^b\,$Instituto de Astrof\'{\i}sica de Canarias, c/ V\'{\i}a 
21: L\'actea s/n, La Laguna, Tenerife, E-38200 Spain}
22: 
23: \altaffiltext{1}{The National Center for Atmospheric Research is
24: sponsored by the National Science Foundation.}
25: 
26: 
27: \begin{abstract}
28: For about a decade, spectro-polarimetry of \ion{He}{1} $\lambda$10830 
29: has been applied to the magnetic diagnostics of the solar chromosphere. 
30: This resonance line is very versatile, as it is visible both on disk 
31: and in off-limb structures, and it has a good sensitivity to both the 
32: weak-field Hanle effect and the strong-field Zeeman effect. Recent 
33: observations of an active-region filament showed that the linear 
34: polarization was dominated by the transverse Zeeman effect, with very 
35: little or no hint of scattering polarization. This is surprising, 
36: since the \ion{He}{1} levels should be significantly polarized
37: in a conventional scattering scenario.
38: To explain the observed level of atomic depolarization by collisional 
39: or radiative processes, one must invoke plasma densities larger by 
40: several orders of magnitude 
41: than currently known values for prominences. We show that
42: such depolarization can be explained quite naturally by the presence 
43: of an unresolved, highly entangled magnetic field, which averages to
44: give the ordered field inferred from spectro-polarimetric data, over the 
45: typical temporal and spatial scales of the observations. We present 
46: a modeling of the polarized \ion{He}{1} $\lambda$10830 in this 
47: scenario, and discuss its implications for the magnetic diagnostics 
48: of prominences and spicules, and for the general study 
49: of unresolved magnetic field distributions in the solar atmosphere.
50: \end{abstract}
51: 
52: \subjectheadings{Sun: chromosphere -- Sun: prominences -- Sun: magnetic
53: fields -- line: profiles -- polarimetry}
54: 
55: \maketitle
56: 
57: Magnetic diagnostics of the chromosphere, and in particular 
58: of prominences and spicules, is receiving increasing attention and 
59: motivation from the solar community. There is in fact a growing 
60: agreement that mapping the magnetic field in these critical 
61: regions of the solar atmosphere is fundamental for our understanding 
62: of processes like coronal heating, the acceleration of the solar 
63: wind, and the release of coronal mass ejections, that have 
64: a direct influence on the heliosphere and on the associated phenomena 
65: of space weather. Despite the strong demand for these critical
66: observations, there have been relatively few attempts to 
67: measure magnetic fields in the chromosphere and corona. This is 
68: mainly due to the heavy science requirements imposed on 
69: spectro-polarimetric instrumentation by this type of observations, and
70: to the intrinsic difficulty of the inversion and interpretation of
71: scattering polarization data.
72: 
73: Spectral line polarization is produced when symmetry-breaking 
74: processes occur in the interaction of radiation with
75: matter. These can be due to the presence of external magnetic or
76: electric fields (Zeeman and Stark effects), or an anisotropy in the
77: excitation of the atoms (by radiation or particles) leading to scattering 
78: polarization. We can expect both types of processes to be always 
79: present in the solar atmosphere. However, their corresponding
80: polarization signals differ significantly. In particular, the presence 
81: of unresolved fields affects those signals in characteristic ways. 
82: For example, for a completely random distribution of magnetic fields
83: within the resolution element of the observations, the 
84: polarization by the Zeeman effect must vanish in the mean, whereas 
85: the scattering polarization is only reduced in amplitude by a 
86: characteristic factor of 1/5 with respect to the zero-field case. 
87: When both processes occur, and the random magnetic field has a 
88: non-zero mean, the observed signal will be a complex mix 
89: of Zeeman effect and scattering polarization, with a significant
90: depolarization contributed by the random part of the magnetic field.
91: 
92: Recent observations of the \ion{He}{1} multiplet at 1083\,nm in an 
93: active-region (AR) filament \citep{Ku09} have shown that the 
94: linear polarization was dominated by the transverse Zeeman effect, 
95: corresponding to magnetic fields in the 500\,G--1000\,G range. 
96: The modeling of the forward-scattered radiation of \ion{He}{1} 
97: $\lambda$10830, under the assumption that the filament plasma is 
98: illuminated by the underlying photosphere, indicates instead that 
99: the linear polarization should be significantly affected by the 
100: atomic alignment of the upper term $\rm 2\,{}^3P$, induced by 
101: radiation anisotropy, even in the presence of field strengths of 
102: the order of $10^3$\,G.
103: A basic assumption of the Zeeman-effect model is that the atomic 
104: levels of \ion{He}{1} are ``naturally'' populated, and therefore 
105: completely depolarized. In contrast, in the scattering-polarization
106: model, which includes atomic polarization, we were forced to 
107: introduce an ad-hoc reduction factor for the radiation anisotropy 
108: (varying between 0 and 1) in order to reproduce the observed 
109: depolarization of the \ion{He}{1} levels. Obviously both 
110: models lack a physical basis for such depolarization, which is 
111: the question taken up in this Letter.
112: 
113: 
114: 
115: The characteristic rate, $\gamma$, at which atomic polarization 
116: is generated in the levels of a spectral line by resonance 
117: scattering, is inversely proportional to the radiative lifetimes of 
118: the levels. In particular, for \ion{He}{1} $\lambda$10830,
119: $\gamma_u\sim A_{ul}\approx 10^7\rm\,s^{-1}$ and $\gamma_l\sim B_{lu}J$, 
120: where $A_{ul}$ and $B_{lu}$ are the Einstein coefficients, 
121: respectively for spontaneous emission and absorption, between the upper 
122: ($u$) and lower ($l$) levels, and $J$ is the average intensity of 
123: the radiation field. If we assume
124: $J\approx (1/2)\,B_\nu(T{=}5800\,\rm K)$, then
125: $\gamma_l\approx(3/20)\,\gamma_u$. 
126: 
127: Once created, atomic polarization is not easily destroyed, 
128: unless competing processes (e.g., external fields, collisions, radiative 
129: ionization and recombination) modify the polarization of the 
130: atomic levels at a much higher rate than both $\gamma_l$ and 
131: $\gamma_u$. A strong magnetic field (i.e., with Larmor frequency
132: $\nu_{\rm L}\gg\gamma_{l,u}$) can depolarize 
133: atomic levels (through the Hanle effect) very efficiently, 
134: although the degree of
135: residual polarization strongly depends on the geometry of the field.
136: For example, scattering polarization is completely destroyed by 
137: a strong magnetic field with inclination 
138: $\vartheta_B=\arccos1/\sqrt{3}\approx 54.7^\circ$
139: from the direction of illumination. This effect may be invoked 
140: to explain the AR filament observations described in \cite{Ku09}. 
141: However, those authors report profiles showing strong atomic 
142: depolarization even for nearly horizontal magnetic fields.
143: 
144: \begin{figure}[!t]
145: \centering
146: \includegraphics[height=3in]{f1.eps}
147: \caption{\label{fig:grotrian}
148: Model atom for the statistical equilibrium of the triplet
149: species of \ion{He}{1}. The lowest five terms constitute the system of
150: bound-bound transitions. The model is extended to include 
151: photo-ionization and recombination involving the $\rm {}^3P$ states, 
152: by adding a fictitious bound state $\rm {}^3S_1$ corresponding to 
153: the ground state of \ion{He}{2}.}
154: \end{figure}
155: 
156: Atomic collisions with an isotropic distribution of perturbers
157: (neutral or charged) are a possible mechanism of atomic
158: depolarization. 1) Inelastic collisions with electrons are not
159: significant for the statistical equilibrium of \ion{He}{1} at
160: the typical densities of prominences 
161: ($N\sim10^{10}$--$10^{11}\,\rm cm^{-3}$). 
162: The observation of scattering polarization at the limb in these 
163: structures also points to the fact that atomic excitation is dominated 
164: by resonance scattering rather than thermal processes. 
165: It is possible 
166: that AR filaments may be significantly denser than quiescent prominences,
167: leading to a larger contribution of thermal processes. However,
168: following \cite{vR62}, collisional excitation of 
169: \ion{He}{1} $\lambda$10830 by electrons would require 
170: $n_{\rm e}\approx 4\times10^{13}\,\rm cm^{-3}$, in order to be comparable 
171: with radiative excitation. This would imply gas densities 3 to 4 
172: orders of magnitude larger than currently known values in quiescent 
173: prominences \citep[e.g.,][]{TH95} for
174: electron collisions to dominate the formation of \ion{He}{1}.
175: 2) We can estimate the rate of elastic 
176: collisions of \ion{He}{1} atoms in the $\rm {}^3P$ state with 
177: neutral hydrogen, using 
178: $\gamma_{\rm H}\sim\langle\sigma v\rangle\,n_{\rm H}\approx 8\times
179: 10^{-9}\,n_{\rm H}\,\rm cm^3\,s^{-1}$ \citep{LT71}. 
180: Hence, to efficiently depolarize the upper levels of \ion{He}{1} 
181: $\lambda10830$ by collisions with neutral hydrogen it should be
182: $n_{\rm H}\gg 10^{15}\,\rm cm^{-3}$.
183: 3) Elastic collisions with electrons have received little
184: attention in the literature. \cite{Hi88} report experimental
185: cross-sections at $T=2000$\,K for \ion{Ne}{1} in the $2p_2$ level.
186: If we extend their results to the case of \ion{He}{1} at
187: plasma temperatures $T\sim 10\,000$\,K, we find 
188: $\gamma_{\rm e}^{\rm el.}\sim 10^{-6}\,n_{\rm e}\,\rm cm^3\,s^{-1}$. 
189: Efficient depolarization of the upper state of \ion{He}{1} 
190: $\lambda10830$ would then imply $n_{\rm e}\gg 10^{13}\,\rm cm^{-3}$, 
191: at which line excitation by thermal electrons should 
192: dominate over radiative excitation (see above).
193: In conclusion, collisional depolarization of the \ion{He}{1} levels
194: seems to consistently require plasma densities that are exceptionally 
195: large for a typical prominence. On the other hand, the densities of 
196: AR filaments are poorly known, and the possibility of such high
197: plasma densities merits further investigation.
198: 
199: \begin{figure}[!t]
200: \centering
201: \includegraphics[width=\hsize]{f2.eps}
202: \figcaption{\label{fig:Xsect}
203: Photo-ionization rates for the lowest $\rm {}^3P$ terms (left) and 
204: $\rm {}^3S$ terms (right) of neutral helium, as a function 
205: of the temperature of a Planckian illumination. The thick lines
206: correspond to the $\rm 2\,{}^3X$ terms, and the thin lines to the
207: $\rm 3\,{}^3X$ terms.}
208: \end{figure}
209: 
210: We consider next the role of radiative processes like 
211: photo-ionization and recombination for the depolarization of 
212: \ion{He}{1} levels.
213: Figure~\ref{fig:grotrian} illustrates the atomic model commonly 
214: employed for the study of resonance scattering polarization in 
215: \ion{He}{1} lines (e.g., \citealt{LA82,LC02,AT08}). This model has 
216: been successfully applied to interpret the Stokes profiles 
217: of the D$_3$ line at 587.6\,nm in solar prominences 
218: \citep{BS78,LA82,CA03}, and in more recent years to the 
219: magnetic diagnostics of quiescent prominences and filaments observed 
220: at 1083\,nm \citep{TB02,Me06}. 
221: The system of bound-bound transitions involves the 
222: lowest five terms of the triplet species of \ion{He}{1}. In order to 
223: include photo-ioniziation and recombination in this model, we added a 
224: fictitious bound state, $\rm {}^3S_1$, corresponding to the
225: ground level of \ion{He}{2} plus a free electron (with spin parallel to
226: that of the \ion{He}{2} ion). 
227: The depopulation rate by photo-ionization is given by
228: $\gamma_{\rm pi}=\int_{\chi/h}^\infty d\nu\;
229: 	(4\pi a_\nu/h\nu)\,J(\nu)$,
230: where $J(\nu)$ is the intensity of the ionizing radiation 
231: averaged over the unit sphere, and $\chi$ is the ionization potential.
232: The ionization cross-section, $a_\nu$, for various atomic terms of
233: \ion{He}{1} is tabulated by the TOPbase project \citep[e.g.,][]{BA05}.
234: In the case of the $\rm 2\,{}^3P$ term, the frequency range over which
235: the cross-section is tabulated goes from the ionization threshold up 
236: to $\sim 3.7$\,nm.
237: Unfortunately, the mean intensity of the solar radiation below the 
238: ionization wavelength of 342.2\,nm is not well known.
239: UV observations of the solar atmosphere have traditionally focused on 
240: limited spectral ranges of particular interest. More complete
241: datasets, covering the whole UV spectrum down to 
242: $\lambda\sim 1$\,nm \citep{EW03}, are integrated over the solar
243: disk, and do not distinguish the output of the quiet Sun from that of 
244: active regions (although they account for variability of the irradiance
245: during the solar cycle).
246: Figure~\ref{fig:Xsect} shows the photo-ionization rate,
247: $\gamma_{\rm pi}$, of the \ion{He}{1} terms as a function of 
248: the effective (Planckian) temperature of the plasma contributing the
249: ionizing radiation. We see that 
250: $\gamma_{\rm pi}\sim\gamma_u$
251: at effective temperatures of $10^5$\,K or larger. On the other
252: hand, the effective temperature of the UV emitting solar plasma, as 
253: estimated by the integrated number of UV photons detected above the 
254: Earth's atmosphere \citep{EW03}, is smaller by at least one order of 
255: magnitude. Therefore the ionizing UV radiation cannot play a 
256: major role in the depolarization of the \ion{He}{1} $\lambda$10830 
257: levels.
258: 
259: 
260: 
261: The depolarizing mechanisms considered above seem unable to 
262: efficiently destroy atomic polarization generated by resonance 
263: scattering, unless one can accept a radical revision of plasma density
264: estimates in AR filaments. The possibility remains that in the 
265: AR filament described by \cite{Ku09} atomic polarization was not
266: generated at all. We considered whether the center-to-limb 
267: variation (CLV) of the solar atmosphere, which dominates the 
268: anisotropy of the illuminating radiation at low
269: heights, may be significantly different in an active region, because 
270: of enhanced lateral illumination coming from plages and/or a brighter 
271: transition region. However, the few measurements of the CLV in active
272: regions \citep[e.g.,][]{SC02} covering the near-infrared solar
273: spectrum do not seem to support this possibility.
274: 
275: 
276: \cite{TA07} have proposed that the radiation inside an optically 
277: thick filament could be nearly isotropic, and
278: therefore unable to induce atomic polarization.
279: By modeling the filament as a free-standing, plane-parallel slab
280: of optical thickness $\tau\sim 1$ and constant source function, $S$,
281: illuminated by the underlying photosphere with intensity $I_0$, 
282: they find that the radiation anisotropy inside the slab may be 
283: significantly reduced and even become negative (because the 
284: illumination is predominantly horizontal). In particular, 
285: when $S\approx I_0$, the anisotropy inside the filament vanishes.
286: However, if the formation of \ion{He}{1} $\lambda10830$
287: is dominated by scattering 
288: ($\gamma_{\rm e}^{\rm inel.}/\gamma_u\sim 0.002$ for 
289: $n_{\rm e}\sim10^{11}\,\rm cm^{-3}$ typical of prominences),
290: the source function must be modeled accordingly. For example, 
291: simply assuming 
292: $S\approx\oint (\mathbf{d\Omega}/4\pi)\,p(\cos\Theta)\,I$,
293: where $p(\cos\Theta)$ is the Rayleigh phase function, the anisotropy
294: rises sharply from about 40\% of its optically thin limit close to 
295: the lower boundary, to well above the optically thin limit at the 
296: free upper boundary. Thus, the anisotropy inside the slab does fall 
297: below the optically thin limit, but complete isotropy is never 
298: attained. Therefore, although a careful radiative transfer treatment
299: of this problem is important, the atomic depolarization of \ion{He}{1} 
300: may not depend only on that. Moreover, the inversions 
301: presented by \cite{Ku09} indicated that the average optical depth of 
302: the filament was sizably smaller than 1, further reducing the 
303: depolarizing role of radiative transfer.
304: 
305: The preceeding analysis motivated us to consider a mechanism of 
306: depolarization not commonly included in polarimetric diagnostics, 
307: namely, that the depolarization of the
308: \ion{He}{1} levels is due to the presence of a significant random 
309: component in the magnetic field inferred from
310: spectro-polarimetric observations.  Both observation and theory support 
311: the idea that solar magnetic fields, even in the quasi-steady state, 
312: are chaotic and dynamic on small spatial and temporal scales.
313: The quiescent-prominence movies made by Hinode/SOT at unprecedented
314: high temporal and spatial resolutions have revealed the extreme 
315: complexity and rapid evolution of filamentary prominence plasma
316: \citep{Be08}.  The long lifetime of a quiescent prominence, days to 
317: weeks, implies a large-scale, stable magnetic topology, but which is 
318: separate from its significantly entangled
319: and rapidly evolving structures at subarsecond scales. MHD waves and 
320: oscillations, together with rising and descending
321: plumes, develop ceaselessly in the macroscopic prominence structure.
322: Although the AR filament described by
323: \cite{Ku09} is a class of prominence distinct from the quiescent 
324: prominences, we may expect the average field inferred from observation 
325: to be co-existing with a small-scale, random component of a comparable 
326: field intensity.  In the extremely low-$\beta$, electrically highly 
327: conducting environment of the AR atmosphere above the photosphere, 
328: current sheets are expected to form densely within any volume 
329: pervaded by a magnetic field endowed with a complex topology
330: \citep{Pa94,JL09}. As these sheets dissipate to heat the atmosphere 
331: quiescently, new sheets form without necessarily producing a major 
332: flare.  The small-scale and rapidly evolving fields associated with a 
333: complexity of current sheets are naturally of an intensity comparable 
334: to the mean field, and could be the origin of the quasi-random 
335: magnetic field we consider here.
336: 
337: 
338: \begin{figure}[!t]
339: \centering
340: \includegraphics[width=\hsize]{f3.eps}
341: \caption{\label{fig:model}
342: Quasi-random magnetic field model (right), obtained by
343: adding a perfectly isotropic (random) field, 
344: $\mathbf{B}_{\rm iso}$, to the inferred mean field, 
345: $\langle\mathbf{B}\rangle$ (left).}
346: \end{figure}
347: 
348: \begin{figure*}[!t]
349: \centering
350: \includegraphics[width=.495\hsize]{f4a.eps}\kern 5pt
351: \includegraphics[width=.495\hsize]{f4b.eps}
352: \caption{\label{fig:Stokes}
353: Examples of Stokes profiles of \ion{He}{1} $\lambda$10830 observed 
354: on the disk, in the presence of a nearly horizontal magnetic field 
355: plus a completely isotropic field of various strengths. 
356: \textit{Left:} $\langle B\rangle=700$\,G, 
357: $B_{\rm iso}=0$\,G (continuous curve), 
358: 200\,G (dotted curve), 
359: 500\,G (dashed curve), 
360: 1000\,G (dashed-dotted curve). \textit{Right:}
361: $\langle B\rangle=20$\,G,
362: $B_{\rm iso}=0$\,G (continuous curve), 
363: 10\,G (dotted curve), 
364: 20\,G (dashed curve), 
365: 100\,G (dashed-dotted curve).}
366: \end{figure*}
367: 
368: Following this picture, we have calculated the emergent Stokes
369: profiles from a homogeneous slab, assuming that the \ion{He}{1} atoms
370: are subject to the mean field inferred from the observations plus a 
371: completely random (hence, isotropic) field of a given strength. 
372: This is also equivalent to assuming a specific 
373: non-isotropic distribution of magnetic fields with varying strength, 
374: resulting exactly in the observed mean field (Fig.~\ref{fig:model}). 
375: It can be argued that this model is very likely an oversimplification 
376: of the real, quasi-random field occurring in prominences. However,
377: its main purpose is to illustrate in the simplest way the depolarizing 
378: effect of such a field. There is no difficulty in principle in calculating 
379: the same effect also for more complicated field distributions.
380: Figure~\ref{fig:Stokes} (left) shows the emergent Stokes profiles for such a 
381: distribution of magnetic fields, resulting in an almost horizontal 
382: mean field of 700\,G, similar to those described by \cite{Ku09}. 
383: The four overplotted profiles illustrate the depolarizing effect induced 
384: by the inclusion of an isotropic field of 200, 500, and 1000\,G, as well 
385: as the case where only the mean field is present. The level of atomic 
386: depolarization inferred from the observations corresponds to an 
387: isotropic field between 500 and 1000\,G, in agreement with the 
388: proposed picture of a very entangled, strong field occurring in the 
389: filament plasma.
390: Figure~\ref{fig:Stokes} (right) also shows the expected effects 
391: of atomic depolarization in the presence of a quasi-random field 
392: averaging to a nearly horizontal, ordered magnetic field with 
393: a strength of 20\,G, typical of quiescent prominences and 
394: spicules. Especially at such small field strengths, the inference
395: of the magnetic field by the Hanle effect becomes very sensitive to 
396: the presence of additional atomic depolarizing processes.
397: 
398: The successful explanation of atomic depolarization by a quasi-random
399: magnetic field stresses the importance to include unresolved fields in
400: the polarimetric diagnostic investigation of the solar atmosphere. The 
401: depolarization effect 
402: of a completely turbulent magnetic field has been studied in detail 
403: \citep[e.g.,][]{St94,LL04}, and has also found many applications in the 
404: study of 
405: quiet-Sun magnetism \citep[see, e.g., the review by][]{TB06}. 
406: In contrast, to our knowledge, no general account has been given in
407: the past of similar depolarization effects produced by a field that 
408: is significantly entangled at scales below the temporal and spatial 
409: resolutions of spectro-polarimetric observations, but which averages 
410: at a finite, mean field when integrated over those scales.
411: Such mean field, which determines the large-scale evolution of the 
412: plasma, remains within the grasp of the traditional diagnostics based 
413: on the Zeeman and Hanle effects. On the other hand, the signature of 
414: atomic depolarization---like in the observations described by 
415: \cite{Ku09}---finds a very simple and direct interpretation in 
416: terms of the strength of a completely random field unresolved 
417: to the observations. This scenario provides a very appealing 
418: interpretational framework for the highly dynamical events 
419: observed at high temporal and spatial resolution with Hinode/SOT, 
420: e.g., in quiescent prominences and spicules \citep{Be08}, thus 
421: opening a new path for the polarimetric study of highly structured 
422: plasmas. 
423: 
424: Of course, this new insight comes at the cost of introducing 
425: an additional degree of freedom in the magnetic diagnostics of 
426: chromospheric fields. For this reason, we strongly advocate that
427: future research programs need to take full advantage of multi-line 
428: spectro-polarimetry, e.g., through simultaneous observations of 
429: \ion{He}{1} $\lambda$10830 and D$_3$ in prominences and spicules,
430: or \ion{He}{1} $\lambda$10830 and H$\alpha$ $\lambda$6563 in 
431: filaments and the chromosphere. In fact, such capability is already 
432: been pursued at the French-Italian solar telescope TH\'EMIS, and 
433: will be the normal mode of operation of HAO's Promincence Magnetometer 
434: \citep[ProMag;][]{El08}, soon to be deployed. We expect that this 
435: type of multi-line diagnostics of chromospheric fields will fully 
436: come to fruition with the large solar facilities of the next 
437: generation, like ATST, EST, and COSMO.
438: 
439: \acknowledgments We thank R.~Centeno Elliott (HAO) and
440: V.~Mart\'{\i}nex Pillet (IAC) for useful comments. RMS acknowledges 
441: support by the Spanish MCYT through project AYA2007-63881.
442: 
443: \begin{thebibliography}{}
444: 
445: 
446: \bibitem[\protect\citeauthoryear{Asensio Ramos, Trujillo
447: Bueno, \& Landi Degl'Innocenti}{2008}]{AT08}
448: Asensio Ramos, A., Trujillo Bueno, J., \& Landi Degl'Innocenti, E.~2008, 
449: \apj, 683, 542
450: 
451: \bibitem[\protect\citeauthoryear{Bauman et al.}{2005}]{BA05}
452: Bauman, R.~P., Porter, R.~L., Ferland, G.~J., \& MacAdam, K.~B.~2005,
453: \apj, 628, 541
454: 
455: \bibitem[\protect\citeauthoryear{Berger et al.}{2008}]{Be08}
456: Berger, T.~E., Shine, R.~A., Slater, G.~L., et al.~2008, \apj, 676, L89
457: 
458: \bibitem[\protect\citeauthoryear{Bommier \& Sahal-Br\'echot}{1978}]{BS78}
459: Bommier, V., \& Sahal-Br\'echot, S.~1978, \aap, 69, 57
460: 
461: \bibitem[\protect\citeauthoryear{Casini et al.}{2003}]{CA03}
462: Casini, R., L\'opez Ariste, A., Tomczyk, S., \& Lites, B.~W.~2003, 
463: \apj, 598, 67L
464: 
465: \bibitem[\protect\citeauthoryear{Elmore et al.}{2008}]{El08}
466: Elmore, D.~F., Casini, R., Card, G.~L., et al.~2008, SPIE, 7014, 39
467: 
468: \bibitem[\protect\citeauthoryear{Eparvier \& Woods}{2003}]{EW03}
469: Eparvier, F.~G., \& Woods, T.~N.~2003, in ISCS Symposium, ESA SP-535, 209
470: 
471: 
472: 
473: 
474: \bibitem[\protect\citeauthoryear{Hirabayashi et al.}{1988}]{Hi88}
475: Hirabayashi, A., Nambu, Y., Hasuo, M., \& Fujimoto, T.~1988, \pra, 37,
476: 83
477: 
478: \bibitem[\protect\citeauthoryear{Janse \& Low}{2009}]{JL09}
479: Janse, \AA.~M., \& Low, B.~C.~2009, \apj\ (in press)
480: 
481: \bibitem[\protect\citeauthoryear{Kuckein et al.}{2009}]{Ku09}
482: Kuckein, C., Centeno Elliott, R., Mart\'inez Pillet, V., et al.~2009 
483: (in preparation)
484: 
485: \bibitem[\protect\citeauthoryear{Lamb \& Ter Haar}{1971}]{LT71}
486: Lamb, F.~K., \& Ter Haar, D.~1971, Phys.\ Rep., 2, 253
487: 
488: \bibitem[\protect\citeauthoryear{Landi Degl'Innocenti}{1982}]{LA82}
489: Landi Degl'Innocenti, E.~1982, \solphys, 79, 291
490: 
491: \bibitem[\protect\citeauthoryear{Landi Degl'Innocenti \&
492: Landolfi}{2004}]{LL04}
493: Landi Degl'Innocenti, E., \& Landolfi, M.~2004, Polarization in
494: Spectral Lines (Dordrecht: Kluwer)
495: 
496: \bibitem[\protect\citeauthoryear{L\'opez Ariste \& Casini}{2002}]{LC02}
497: L\'opez Ariste, A., \& Casini, R.~2002, \apj, 575, 529
498: 
499: \bibitem[\protect\citeauthoryear{Merenda et al.}{2006}]{Me06}
500: Merenda, L., Trujillo Bueno, J., Landi Degl'Innocenti, E., \& Collados,
501: M.~2006, \apj, 642, 554
502: 
503: \bibitem[\protect\citeauthoryear{Parker}{1994}]{Pa94}
504: Parker, E.~N.~1994, Spontaneous Current Sheets in Magnetic Fields
505: (Oxford: Oxford University)
506: 
507: \bibitem[\protect\citeauthoryear{S\'anchez Cuberes et al.}{2002}]{SC02}
508: S\'anchez Cuberes, M., V\'azquez, M., Bonet, J.~A., \&
509: Sobotka, M.~2002, \apj, 570, 886
510: 
511: \bibitem[\protect\citeauthoryear{Stenflo}{1994}]{St94}
512: Stenflo, J.~O.~1994, Solar Magneti Fields (Dordrecht: Kluwer)
513: 
514: \bibitem[\protect\citeauthoryear{Tandberg-Hannsen}{1995}]{TH95}
515: Tandberg-Hanssen, E.~1995, The Nature of Solar Prominences (Dordrecht:
516: Kluwer)
517: 
518: \bibitem[\protect\citeauthoryear{Trujillo Bueno et al.}{2002}]{TB02}
519: Trujillo Bueno, J., Landi Degl'Innocenti, E., Collados, M., 
520: Merenda, L., \& Manso Sainz, R.~2002, \nat, 415, 403
521: 
522: \bibitem[\protect\citeauthoryear{Trujillo Bueno \& Asensio Ramos}{2007}]{TA07}
523: Trujillo Bueno, J., \& Asensio Ramos, A.~2007, \apj, 655, 642
524: 
525: \bibitem[\protect\citeauthoryear{Trujillo Bueno, Asensio Ramos, \&
526: Shchukina}{2006}]{TB06}
527: Trujillo Bueno, J., Asensio Ramos, A., \& Shchukina, N.~2007, in ASP
528: Conf.\ Ser.\ Vol.\ 358, Solar Polarization 4, ed.\ R.~Casini \&
529: B.~W.~Lites (San Francisco: ASP), 269
530: 
531: \bibitem[\protect\citeauthoryear{van Regemorter}{1962}]{vR62}
532: van Regemorter, H.~1962, \apj, 136, 906
533: 
534: 
535: 
536: 
537: 
538: 
539: 
540: 
541: 
542: 
543: 
544: \end{thebibliography}
545: 
546: \end{document}
547: 
548: