1:
2: \documentclass[useAMS,usenatbib]{mn2e}
3: \usepackage{amsmath,amssymb}
4: %
5: \def\rmmat#1{{\hbox{\rm #1}}}
6: \def\rmscr#1{\rmmat{\scriptsize #1}}
7: % Additional packages needed for graphics, alignment and math fonts.
8:
9: %\usepackage[number,sort,compress]{natbib}
10: \usepackage{graphicx} % Include figure files in eps format.
11: \usepackage{dcolumn} % Align table columns on decimal point.
12: \usepackage{bm} % bold math.
13: \usepackage{psfrag}
14: \newcommand{\aap}{Astron. Astrophys.}
15: \newcommand{\apj}{Astrophys. J}
16:
17: \begin{document}
18:
19: \title[Magnetospheric Birefringence]{Magnetospheric Birefringence Induces Polarization Signatures in Neutron-Star Spectra}
20:
21: \author[R. M. Shannon and J. S. Heyl]{R. M. Shannon\thanks{Current Address: Department of Astronomy,
22: Cornell University, Ithaca, New York, United States, 14853; email: ryans@astro.cornell.edu}
23: and J. S. Heyl\thanks{email: heyl@phas.ubc.ca; Canada Research Chair}\\
24: Department of Physics and Astronomy, University of British Columbia, Vancouver, British Columbia, Canada, V6T 1Z1}
25:
26: \date{Accepted 2006 February 16. Received 2006 February 15; in original form 2006 January 19 }
27:
28: \pagerange{\pageref{firstpage}--\pageref{lastpage}} \pubyear{2006}
29:
30: \maketitle
31: \label{firstpage}
32:
33:
34: \begin{abstract}
35: We study the propagation of polarization light through the
36: magnetosphere of neutron stars. At intermediate frequencies (the
37: optical through the infrared), both the birefringence induced by the
38: plasma and by quantumelectrodynamics influence the observation
39: polarization of radiation from the surface of the neutron star.
40: Because these two processes compete in this regime, we find that
41: polarization observations can constrain the properties of the
42: neutron-star magnetosphere, specifically the total charge density. We
43: calculate both the phase-resolved and the phase-averaged polarization
44: signatures induced by magnetospheric birefringence.
45: \end{abstract}
46:
47: \begin{keywords}
48: stars : neutron -- magnetic fields -- radiation mechanisms : general
49: \end{keywords}
50:
51: \section{Introduction}
52:
53: The propagation of polarized radio waves away from neutron stars (NSs)
54: has been well understood for many years \citep{Barn86}. At this
55: energy scale, it is reasonable to neglect the coupling of the light to
56: vacuum-based QED effects. Lai and Ho
57: (2003ab)\nocite{Lai:2003,Lai:2003b} have discussed the mechanisms by
58: which polarized X-rays are affected by the atmospheres of NSs. X-ray
59: photons couple the greatest in the atmosphere as the plasma densities
60: and field strengths are sufficiently high. Lai and Ho included
61: vacuum-based effects in their analysis. Similar effects exist at
62: IR/optical wavelengths when propagation of the waves through the
63: magnetosphere (possessing a lower-density plasma and lower strength
64: magnetic field) is considered. This paper presents a model and
65: results that show that there is a wavelength-dependent polarization
66: signature in the IR/optical wavelength as a result of interactions
67: between radiation and the magnetosphere.
68:
69: For most radio pulsars emission from the magnetosphere itself rather
70: than the surface is thought to dominate the optical and near infrared
71: emission. However, several radio pulsars exhibit optical radiation
72: from the surface ({\em e.g.} PSR~B0656+14, PSR~B0950+08 and
73: Geminga) \citep[\protect{\em e.g.}][]{2004A&A...417.1017Z}. Radio pulsars are not the only
74: members of the neutron-star menagerie. Tbe optical and near-infrared
75: emission from the radio quiet neutron stars such as RX~J1856.5-3754
76: and RX~J0720.4-3125\citep[\protect{\em e.g.}][]{2002ApJ...564..981P,2003A&A...408..323M} most
77: likely comes from the surface, and possibly the emission from
78: anomalous x-ray pulsars emerges from a surface layer heated by
79: magnetospheric currents \citep{2004astro.ph..8538T}. With these
80: objects in mind let us proceed.
81:
82: To make progress we will make several simplifying assumptions.
83: First, we assume that the charge carriers both positive and negative
84: have the mass of an electron and that the plasma is cold. A pair
85: plasma is necessarily hot. Second, we assume that the Goldreich-Julian
86: model is correct. The Goldreich-Julian picture describes the
87: equilibrium charge density for a rotating neutron star with its
88: magnetic pole aligned (or anti-aligned) with the spin axis. Here we
89: will assume that the two axes are not necessarily aligned (especially
90: in \S~\ref{sec:phase-integrations}). Finally, the Goldreich-Julian
91: equilibrium even in its realm of applicability is unstable
92: \citep[see][]{2004AdSpR..33..542M}. In spite of these deficiencies,
93: the Goldreich-Julian model provides a well-understood starting point
94: to demonstrate that observations of polarization of the surface
95: radiation from neutron stars may provide a window to diagnose the
96: properties of the plasma and to encourage further research to probe
97: more realistic models.
98:
99: \section{Calculations : Polarization Evolution}
100:
101: \citet{Kubo:1983} outline how the polarization of light is
102: affected by a birefringent and dichroic medium:
103: \begin{equation}
104: \frac{d \vec{s}}{dz} = \hat{\Omega} \times \vec{s} + (\hat{T} \times \vec{s}) \times \vec{s}
105: \label{propeqn}.
106: \end{equation}
107: This equation is valid for a medium with dispersion, emission and
108: absorption but not scattering.
109: The polarization is described in terms of normalized Stokes parameters
110: $\vec{s} = \left(s_1,s_2,s_3 \right) =$ (Q/I, U/I, V/I). The vectors
111: $\hat{\Omega}$ and $\hat{T}$ are the
112: birefringent and dichroic
113: vectors. We have extended the expressions of
114: \citet{Heyl:2000} for a permeable medium to include the effects
115: of the plasma:
116: \begin{equation}
117: \hat{\Omega} = \frac{\omega}{2 c \sqrt{\epsilon_0}} \left (\begin{array}{c}
118: (\eta_E \epsilon_B - \epsilon_E \eta_B) \sin^2 \theta \cos 2 \phi \\
119: -(\eta_E \epsilon_B - \epsilon_E \eta_B) \sin^2 \theta \sin 2 \phi\\
120: g \cos \theta ( \epsilon_B + \frac{1}{2}(\eta_B - \epsilon_B)\sin^2 \theta)
121: \end{array} \right)
122: \label{birefringent},
123: \end{equation}
124: where $\omega$ is the angular frequency of the radiation;
125: $\theta$ is the angle between the direction of propagation and
126: the magnetic field; and $\phi$ is the angle between the projection of
127: the magnetic field onto the plane perpendicular to the direction of
128: propagation, and the component of the magnetic field in the direction
129: of the magnetic moment of the NS. $\eta_B$ and $\epsilon_B$
130: are components of the permeability tensor, and $\eta_E$ and
131: $\epsilon_E$ are components of the dielectric tensor, which are, to
132: first order in the fine-structure constant ($\alpha_f$):
133: \begin{eqnarray}
134: \eta_B &=& 1 + \frac{2}{15}\frac{\alpha_f}{\pi} \left(\frac{B}{B_Q}
135: \right)^2, \\
136: \epsilon_B &=& 1 + \frac{2}{45}\frac{\alpha_f}{\pi} \left(\frac{B}{B_Q} \right)^2,\\
137: \eta_E &=& 1 -v + \frac{1}{9}\frac{\alpha_f}{\pi} \left(\frac{B}{B_Q} \right)^2,\\
138: \epsilon_E &=& 1 - \frac{v}{1-u} - \frac{2}{45} \frac{\alpha_f}{\pi}\left(\frac{B}{B_Q} \right)^2,\\
139: g &=& \frac{v \sqrt{u}}{1 - u},\\
140: \epsilon_0 &=& \epsilon_E \epsilon_B \cos ^2 \theta + \frac{1}{2} \sin^2 \theta
141: \left( \eta_E \epsilon_B + \epsilon_B \eta_B \right)
142: \end{eqnarray}
143: for $B \ll B_Q$ and $\hbar \omega \ll m_e c^2$ where $v =
144: \omega_p^2/\omega^2$, and $u = \omega_c^2/\omega^2$.
145: $\omega_c=eB/(m_e c)$ is the cyclotron frequency and
146: \begin{equation}
147: \omega_p =
148: \sqrt {\frac{4\pi e^2n_e}{m_e}} =
149: \sqrt{2 \Omega \omega_c |\cos \theta'|}
150: \end{equation}
151: is the plasma frequency under the assumptions mentioned in the
152: introduction.
153:
154: To get the second equality we have assume the Goldreich-Julian
155: (1969) \nocite{Goldreich:1969} plasma density, $n_{GJ}=(\Omega
156: B\cos\theta')/(2\pi ec)$. $\Omega=2\pi/P$ is the spin frequency of
157: the star and $\theta'$ is the angle between the rotation axis and the
158: local magnetic field. $B$ is the strength of the magnetic field {\it
159: perpendicular to propagation direction} and $B_Q \approx 4.4 \times
160: 10^{13}$~G is the critical QED field. In the range of NSs modeled, it
161: was found that the dichroic vector had little effect on the
162: polarization, and was neglected for subsequent results presented here.
163:
164: A fourth-order Runge-Kutta integrator with variable step size was used
165: to determine the evolution of the polarization vector from the surface
166: of the star to the light cylinder, along a Schwarzschild null
167: geodesic. The polarization vector $\vec{s}$ was initially set to be
168: in the direction of the plasma-dominated birefringent vector, as the
169: initial polarization modes are determined by the plasma-rich
170: atmosphere of the NS.
171:
172: We set the initial conditions by specifying the inclination angle
173: $\alpha$ (not to be confused with the fine structure constant
174: $\alpha_f$) between the optical axis and the magnetic axis, and the
175: angle $\beta$ between the plane containing the optical and magnetic
176: axes and the plane containing the photon trajectory. For all the
177: calculations stellar radii $R = 10^6$ cm and mass $M = 1.4$
178: $M_{\odot}$ were considered. In addition, we assumed that the NS's
179: magnetic field was dipolar. The output of single integration is
180: similar to Figure~3 of \citet{Heyl:2003}. The $s_1$ and $s_2$
181: polarizations change considerably as the photon travels away from the
182: NS. At approximately 30 stellar radii, the polarization decouples.
183: The significance of the decoupling radius is discussed below.
184:
185: \section{Results}
186:
187: \subsection{Net Polarization}
188:
189: The net polarization of the radiation emitted from the NS
190: was determined by integrating over many trajectories on the neutron
191: star image. We sampled the image using regions of equal solid angle
192: using concentric rings about the
193: centre of the NS image. This sampling method could be
194: performed due to the inherent symmetry in the circular NS
195: image. By symmetry, the $s_2$ and $s_3$ components of the
196: polarization are zero. Figure \ref{compare} shows the
197: frequency-dependent polarization signature for a NS with a
198: dipole moment of $\mu=10^{31}$G~cm$^3$. The full-model polarization
199: (solid line) is plasma dominated at low frequencies and vacuum (QED)
200: dominated at higher frequencies. In the transitory region, the total
201: polarization suffers some destructive interference between the two
202: processes -- the total polarization is less than either the vacuum
203: or plasma-dominated cases.
204: \begin{figure}
205: \psfrag{s1}{$s_1$}
206: \psfrag{ 1e+13}{$10^{13}$}
207: \psfrag{ 1e+14}{$10^{14}$}
208: \psfrag{ 1e+15}{$10^{15}$}
209: \psfrag{30m}{$30\mu$m}
210: \psfrag{3m}{$3\mu$m}
211: \psfrag{0.3m}{$0.3\mu$m}
212: \includegraphics[width=7.6cm,height=4.6cm]{13.ps}
213: \caption{\label{compare} Comparison of the three regimes for a neutron
214: star with a one-second spin period and a magnetic dipole moment of
215: $10^{31}$G~cm$^3$, pointing in a direction $60^\circ$ from the line of
216: sight. The solid line represents the model containing both QED and
217: plasma-based effects, the dot-dashed line shows integrations
218: conducted using only plasma effects, and the short-dashed lines
219: represents integrations conducted with only vacuum-based effects.}
220: \end{figure}
221:
222: The location of the minimum in the transitory region can be understood
223: to be the frequency at which the photon decouples from the plasma and vacuum
224: effects at the same time. Using the adiabatic condition,
225: \citet{Heyl:2000} determined the criteria for decoupling,
226: \begin{equation}
227: \left|\hat{\Omega} \left( \frac{1}{|\hat{\Omega}|}\frac{\partial
228: |\hat{\Omega}|}{\partial r} \right)^{-1} \right| \approx 0.5
229: \end{equation}
230:
231: Using this expression, the distance at which decoupling occurs can be
232: solved for in both plasma-dominant and QED-dominant regimes. To first
233: order these expressions are:
234: \begin{equation}
235: r_\rmscr{pl, qed} \propto \mu^{\frac{2}{5}} \omega^{\frac{1}{5}}
236: \times f(\alpha);\,
237: r_\rmscr{pl, plas} \propto \left (\frac{\mu}{P \omega} \right)^{\frac{1}{2}} \times g(\alpha)
238: \end{equation}
239:
240: Here $f(\alpha)$ and $g(\alpha)$ denote the functional dependence of
241: the expressions on the inclination of the star. As the resonance
242: occurs when the photon decouples from both the vacuum and plasma
243: dominated effects at the same time, the critical frequency $\omega_c$
244: and the critical radius $r_c$ at which this resonance occurs are
245: easily calculated:
246: \begin{equation}
247: \omega_c \propto \frac{\mu^{\frac{1}{7}}}{P^{\frac{5}{7}}} \times h(\alpha);\,
248: r_c \propto \frac{\mu^{\frac{3}{7}}}{P^{\frac{1}{7}}} \times j(\alpha)
249: \end{equation}
250: The results of the numerical model were found to be in agreement with
251: these relations.
252:
253: The critical polarization $P_c$ (the polarization at resonance) is
254: related to the to the radius at which photons decouple:
255: $1- P_c \propto \left(R/r_c\right)^{\gamma}$
256: where $0 < \gamma \lesssim 2$.
257:
258: Figure \ref{compB} shows how the polarization signature changes as the
259: magnetic field strength is varied. It is observed that as the magnetic
260: field strength is increased, the strength of the resonance feature
261: decreases, but the critical frequency doesn't change significantly.
262: \begin{figure}
263: \psfrag{s1}{$s_1$}
264: \psfrag{ 1e+13}{$10^{13}$}
265: \psfrag{ 1e+14}{$10^{14}$}
266: \psfrag{ 1e+15}{$10^{15}$}
267: \psfrag{30m}{$30\mu$m}
268: \psfrag{3m}{$3\mu$m}
269: \psfrag{0.3m}{$0.3\mu$m}
270: \includegraphics[width=7.6cm,height=4.6cm]{compB.ps}
271: \caption{\label{compB} Polarization signatures for various magnetic
272: field strengths, with period of 1.0s. The plot shows polarization
273: signatures for increasing magnetic fields, with the lowest line
274: showing a field with a dipole moment of $2 \times 10^{31}$G~cm$^3$,
275: each subsequent line representing a factor of two increase in the
276: magnetic field to the last line, which shows the polarization
277: signature for a magnetic dipole moment of $6.4 \times
278: 10^{32}$G~cm$^3$
279: Above each curve is the magnetic dipole moment in units of
280: $10^{31}$~G~cm$^3$. In each case the dipole points
281: 30$^\circ$ from the line of sight.
282: }
283: \end{figure}
284: Varying the rotational period of the NS does change the
285: polarization signature. The variation is due to the
286: effects of increased plasma density in the magnetosphere. The results
287: seen in Figure \ref{period} show that as the rotational period is
288: changed the frequency of the dominant features are changed, but the
289: overall structure of the signature remains similar.
290: %\begin{figure}
291: %\psfrag{s1}{$s_1$}
292: %\psfrag{ 1e+13}{$10^{13}$}
293: %\psfrag{ 1e+14}{$10^{14}$}
294: %\psfrag{ 1e+15}{$10^{15}$}
295: %\includegraphics[width=7.6cm, height=4.6cm]{period.ps}
296: %\caption{\label{period} Polarization signatures for a dipole moment of
297: %$10^{31}$G cm$^3$ and various periods. The plots shows the
298: %polarization signatures for various NS periods. The lines show,
299: %starting from the most significant minimum in the left periods of 70s, 32s,
300: %20s,10s, 7s, 1s, 0.32s, 0.2s, 0.1s, and 0.07s.}
301: %\end{figure}
302:
303: \begin{figure}
304: \psfrag{s1}{$s_1$}
305: \psfrag{ 1e+13}{$10^{13}$}
306: \psfrag{ 1e+14}{$10^{14}$}
307: \psfrag{ 1e+15}{$10^{15}$}
308: \psfrag{30m}{$30\mu$m}
309: \psfrag{3m}{$3\mu$m}
310: \psfrag{0.3m}{$0.3\mu$m}
311: \includegraphics[width=7.6cm, height=4.6cm]{peri_plas.eps}
312: \caption{\label{peri_plas}\label{period}\label{Plasma2}
313: Polarization signatures for a dipole moment of
314: $10^{31}$G cm$^3$ and various periods. The upper panel shows the
315: polarization signatures for various NS periods. The lines show,
316: starting from the most significant minimum on the left periods of 70s, 32s,
317: 20s,10s, 7s, 3.2s, 2s, 1s, 0.32s, 0.2s, 0.1s, and 0.07s. Some of the
318: minima are labeled with the period.
319: The lower panel gives
320: the results for various total charge densities.
321: The lowest line shows the case in which the total charge
322: density is equal to the net charge density. Each successive line above
323: represents an introduction of an additional component of the total charge
324: density proportional to the net charge density. The second lowest
325: line represents an 25\% excess of charge, and each line
326: above that increases the excess by a factor of two. The minima are
327: labeled with the ratio of the total charge to the net charge.
328: In each case the dipole points
329: 60$^\circ$ from the line of sight. }
330: \end{figure}
331:
332: It is possible that the charge density of the plasma has been
333: underestimated. \citet{Goldreich:1969} estimate only the net charge
334: density. However, it is possible that the total charge density is
335: larger, as electron-positron pairs may exist in the magnetosphere.
336: Extending the equations of \citet{Mezaros:1992} into the formalism
337: presented here, the total charge density can be increased arbitrarily.
338: Figure \ref{Plasma2} shows that the introduction of additional charge
339: shifts the resonance to higher frequencies, and decreases the depth of
340: the feature. Again, the overall shape of the feature remains similar.
341: Note that increasing the charge density is comparable to increasing
342: the rotational rate of the NS. At these frequencies increasing the
343: total charge density is nearly covariant with increasing the spin rate
344: of the star, but the ellipticity of the modes depends only on the net
345: charge density so the results for higher net charge density
346: (increasing $\Omega$) and higher total charge density are slightly
347: different.
348: %\begin{figure}
349: %\psfrag{s1}{$s_1$}
350: %\psfrag{ 1e+13}{$10^{13}$}
351: %\psfrag{ 1e+14}{$10^{14}$}
352: %\psfrag{ 1e+15}{$10^{15}$}
353: %\includegraphics[width=7.6cm,height=4.6cm]{plasma2.ps}
354: %\caption{\label{Plasma2} Spectral signatures for a dipole moment of
355: %$10^{31}$G cm$^3$, period of 1.0s, and various total charge
356: %densities. The lowest line shows the case in which the total charge
357: %density is equal to the net charge density. Each successive line above
358: %represents an introduction of an additional component of the total charge
359: %density proportional to the net charge density. The second lowest
360: %line represents an 25\% excess of charge, and each line
361: %above that increases the excess by a factor of two. }
362: %\end{figure}
363:
364: \subsection{Phase Integrations}
365: \label{sec:phase-integrations}
366:
367: It is currently impractical to measure the instantaneous polarization
368: of an object as faint as a neutron star; therefore it is important to
369: calculate the polarization signature from NS averaged over many spin
370: periods. To do this the polarization must be calculated in a single
371: basis. \citet{Heyl:2003} show how the
372: polarization can be calculated. Define the latitude of the optical axis
373: (observer) above rotational equator is as $i_R$ and the angle
374: between the magnetic axis and rotational axis as $\gamma$.
375: If the observer's polarimeter is aligned so that $s_{O,1}$ is
376: along the rotational equator of the NS and $s_{O,2}$ is defined
377: 45$^\circ$ to it, the components of the polarization can be determined
378: from the calculations in the other bases\footnote[14]{The expression for
379: $s_{O,2}$ in \citet{Heyl:2003} is incorrect by a
380: factor of $\sin i_R$. Note that the quadrature sum of $s_{O,1}$ and
381: $s_{O,2}$ should be unity. This result is corrected here. }
382: \begin{eqnarray}
383: s_{O,1} = s_1 \frac{2 \left( \sin \gamma \sin \phi \right) ^2}{1 -
384: \left( \cos\gamma \sin i_R + cos i_R \sin \gamma \cos \phi \right)^2}
385: -1\, \\
386: s_{O,2} = s_1 \frac{2 \sin \gamma \sin \phi \left(\cos \gamma
387: \cos i_R - \cos \phi \sin \gamma \sin i_R \right)}{1 - \left( \cos
388: \gamma \sin i_R + \cos i_R \sin \gamma \cos \phi \right)^2}
389: \end{eqnarray}
390: The angle $\alpha$ as defined in the previous section can be
391: calculated using the law of cosines.
392: Figure~\ref{Single} shows how the polarization from a star evolves as
393: the phase varies. The quadrature sum of the two terms adds to make
394: $s_\rmscr{tot}$, as expected.
395:
396: \begin{figure}
397: \psfrag{blank}{Polarization}
398: \psfrag{stot}{$s_\rmscr{tot}$}
399: \psfrag{sO1}{$s_{O,1}$}
400: \psfrag{sO2}{$s_{O,2}$}
401: \includegraphics[width=7.6cm,height=4.6cm]{single.ps}
402: \caption{\label{Single} Polarization from Neutron Star as a function
403: of phase for a NS with $\gamma = 30^{o}$ and $i_R = 30^\circ$,
404: for $\omega = 10^{13}$ rad/s and $\mu = 10^{31}$ G cm$^3$. $s_\rmscr{tot}$
405: is the total polarization at the particular phase. $s_{O,1}$ is the
406: component of polarization measured in direction perpendicular and
407: parallel to the rotational equator of the NS, and $s_{O,2}$ is the
408: component 45$^\circ$ to the equator.}
409: \end{figure}
410:
411: The phase-averaged polarization can be found by computing the mean values
412: in $s_{O,1}$ and $s_{O,2}$ with the average value of $s_{O,2}$ being zero by
413: symmetry. The lower panel of Figure~\ref{irg} shows how the signature changes as $i_R$
414: varies. Observe that when $\sin i_R = 1$ (top line) -- the observer is
415: looking directly down on the rotation axis -- there is no observed
416: polarization. However, for observations at lower latitudes there is a
417: significant net polarization and a resonant feature at approximately
418: $3 \times 10^{13}$ Hz. Statistically, half of all NS will
419: have inclination angles such that $ \sin i_R \leq 0.5$. All the
420: models presented in this range show strong features in this range
421: considered.
422:
423: %\begin{figure}[t]
424: %\psfrag{sO1}{$s_{O,1}$}
425: %\psfrag{ 1e+13}{$10^{13}$}
426: %\psfrag{ 1e+14}{$10^{14}$}
427: %\psfrag{ 1e+15}{$10^{15}$}
428: %\includegraphics[width=7.6cm,height=4.6cm]{cgamma.ps}
429: %\caption{\label{Gamma} Polarization Signatures for various $i_R$ with
430: %$\gamma = 30^\circ$. The lines show how the signature changes from
431: %$\sin i_R = 0$ (bottom line) to $\sin i_R = 1.0$ (top line,
432: %indistinguishable from the x-axis) with each above representing an
433: %increase in $\sin i_R$ of 0.1. A dipole moment of $10^{31}$ G cm$^3$
434: %and an NS period of 1.0s were assumed.}
435: %\end{figure}
436:
437: The upper panel of Figure~\ref{irg} shows how the signature changes as
438: $\gamma$ varies. The strongest resonant feature occurs when the
439: magnetic and rotational axes are aligned ($\gamma=0$). However,
440: strong resonant features can be observed for $\cos \gamma \ge 0.5$.
441:
442: %\begin{figure}[t]
443: %\psfrag{sO1}{$s_{O,1}$}
444: %\psfrag{ 1e+13}{$10^{13}$}
445: %\psfrag{ 1e+14}{$10^{14}$}
446: %\psfrag{ 1e+15}{$10^{15}$}
447: %\includegraphics[width=7.6cm,height=4.6cm]{cir.ps}
448: %\caption{\label{ir} Polarization Signatures for various $\gamma$ with
449: %$i_R = 30^\circ$. The lines show how the signature changes from $\cos
450: %\gamma = 1.0 $ (bottom line) to $\cos \gamma = 0$ (top line) with each
451: %above representing an decrement in $\cos \gamma$ of 0.1. A
452: %dipole moment of $10^{31}$G~cm$^3$ and an NS period of 1.0s were
453: %assumed.}
454: %\end{figure}
455:
456:
457: \begin{figure}
458: \psfrag{sO1}{$s_{O,1}$}
459: \psfrag{ 1e+13}{$10^{13}$}
460: \psfrag{ 1e+14}{$10^{14}$}
461: \psfrag{ 1e+15}{$10^{15}$}
462: \psfrag{30m}{$30\mu$m}
463: \psfrag{3m}{$3\mu$m}
464: \psfrag{0.3m}{$0.3\mu$m}
465: \includegraphics[width=7.6cm,height=4.6cm]{cir_gamma.eps}
466: \caption{\label{irg}
467: Phase-integrated polarization signatures.
468: The upper panel shows polarization signatures for various $i_R$ with
469: $\gamma = 30^\circ$.
470: The lines show how the signature changes from
471: $\sin i_R = 0$ (bottom line) to $\sin i_R = 1.0$ (top line,
472: indistinguishable from the x-axis) with each above representing an
473: increase in $\sin i_R$ of 0.1. The curves are labelled with the
474: value of $\sin i_R$.
475: The lower panel shows the results
476: for various $\gamma$ with
477: $i_R = 30^\circ$. The lines show how the signature changes from $\cos
478: \gamma = 1.0 $ (bottom line) to $\cos \gamma = 0$ (top line) with each
479: above representing an decrement in $\cos \gamma$ of 0.1.
480: The curves are labelled with the value of $\cos \gamma$. In both
481: panels a
482: dipole moment of $10^{31}$G~cm$^3$ and an NS period of 1.0s were
483: assumed. }
484: \end{figure}
485:
486:
487: \section{Discussion}
488:
489: Of particular attention is the wavelength at which the resonant
490: feature occurs - the near infrared. As this energy, it would be
491: possible to measure the polarization of the thermal radiation from a
492: NS using a large ground-based telescope or Hubble's NICMOS instrument.
493: This instrument is only sensitive to wavelengths less than 2.1~$\mu$m,
494: so it is possible that a large portion of the behaviour could not be
495: probed. If the total plasma density is significantly larger than the
496: Goldreich-Julian plasma density, the feature moves into near infrared
497: or optical energies, and more observing options become available.
498:
499: This letter presents a model for the evolution of polarized light
500: through the magnetosphere of highly magnetized neutron stars. We
501: found that the polarization has a frequency-dependent
502: signature, where the total polarization decreases significantly in the
503: optical or near IR, dependent on the plasma density near the neutron
504: star and the structure of the magnetic field. Observations of neutron
505: stars in this frequency region may provide useful probes of
506: the properties of neutron-star magnetospheres and elucidate the processes
507: responsible for pulsar emission.
508:
509: \section*{Acknowledgments}
510:
511: The Natural Sciences and Engineering Research Council of Canada
512: supported this work. Correspondence and requests for materials should
513: be addressed to J.S.H.(heyl@phas.ubc.ca).
514:
515:
516: \bibliographystyle{mn2e}
517: \bibliography{thesis}
518:
519: \label{lastpage}
520:
521: \end{document}
522: