1: %% preprint produces a one-column, single-spaced document:
2: \documentclass[12pt,preprint]{aastex}
3: %% manuscript produces a one-column, double-spaced document:
4: % \documentclass[manuscript]{aastex}
5: %% preprint2 produces a double-column, single-spaced document:
6: % \documentclass[preprint2]{aastex}
7:
8: \usepackage{natbib}
9: %\citestyle{aa}
10:
11: %% If you want to create your own macros, you can do so
12: %% using \newcommand. Your macros should appear before
13: %% the \begin{document} command.
14:
15: \slugcomment{Submitted to the Astrophysical Journal}
16:
17: %% If you wish, you may supply running head information,
18: %% although this information may be modified by the editorial
19: %% offices. The left head contains a list of authors,
20: %% usually a maximum of three (otherwise use et al.).
21: %% The right head is a modified title of up to roughly
22: %% 44 characters. Running heads will not print in the
23: %% manuscript style.
24:
25: %\shorttitle{}
26: %\shortauthors{}
27:
28: \begin{document}
29:
30: \title{Near-Infrared, Kilosecond Variability of the Wisps
31: and Jet \\ in the Crab Pulsar Wind Nebula}
32:
33: %% Use \author, \affil, and the \and command to format
34: %% author and affiliation information.
35: %% You can use \email to mark an email address
36: %% anywhere in the paper, not just in the front matter.
37: %% As in the title, you can use \\ to force line breaks.
38:
39: \author{A. Melatos and D. Scheltus}
40: \affil{School of Physics, University of Melbourne,
41: Parkville, VIC 3010, Australia}
42: \email{a.melatos@physics.unimelb.edu.au}
43: \author{M. T. Whiting}
44: \affil{School of Physics, University of New South Wales,
45: Sydney, NSW 2052, Australia}
46: \author{S. S. Eikenberry}
47: \affil{Department of Astronomy, University of Florida,
48: Gainesville, FL 32611, USA}
49: \author{R. W. Romani}
50: \affil{Department of Physics, Stanford University,
51: Stanford, CA 94305, USA}
52: \author{F. Rigaut}
53: \affil{Gemini Observatory, Northern Operations Center,
54: 670 N. A'ohoku Place, Hilo, HI 96720, USA}
55: \author{A. Spitkovsky}
56: \affil{KIPAC, Stanford University, PO Box 20450,
57: Stanford, CA 94309, USA}
58: \author{J. Arons}
59: \affil{Department of Astronomy, 601 Campbell Hall,
60: University of California at Berkeley,
61: Berkeley, CA 94720, USA}
62: \and
63: \author{D. J. B. Payne}
64: \affil{School of Physics, University of Melbourne,
65: Parkville, VIC 3010, Australia}
66:
67: %% Notice that each of these authors has alternate affiliations, which
68: %% are identified by the \altaffilmark after each name. Specify alternate
69: %% affiliation information with \altaffiltext, with one command per each
70: %% affiliation.
71: %\altaffiltext{1}{Visiting Astronomer, Cerro Tololo Inter-American Observatory.
72: %CTIO is operated by AURA, Inc.\ under contract to the National Science
73: %Foundation.}
74:
75: \begin{abstract}
76: We present a time-lapse sequence of 20 near-infrared
77: ($J$- and $K'$-band)
78: snapshots of the central $20\arcsec\times 20\arcsec$
79: of the Crab pulsar wind nebula,
80: taken at subarcsecond resolution with the Hokupa'a/QUIRC
81: adaptive optics camera on the Gemini North Telescope,
82: and sampled at intervals of 10 minutes and 24 hours.
83: It is observed that the equatorial wisps and polar knots
84: in the termination shock of the pulsar wind
85: appear to fluctuate in brightness on kilosecond time-scales.
86: Maximum flux variations of
87: $\pm 24\pm 4$ and $\pm 14\pm4$
88: per cent relative to the mean (in $1.2\,{\rm ks}$)
89: are measured for the wisps and knots respectively,
90: with greatest statistical significance in $J$ band
91: where the nebula background is less prominent.
92: The $J$ and $K'$ flux densities imply different near-infrared spectra
93: for the nonthermal continuum emission from the wisps and
94: outermost polar knot (`sprite'),
95: giving $F_\nu \propto \nu^{-0.56\pm0.12}$ and
96: $F_\nu \propto \nu^{-0.21\pm0.13}$ respectively.
97: The data are compared with existing optical and UV photometry
98: and applied to constrain theories of the
99: variability of the wisps (relativistic ion-cyclotron
100: instability) and knots (relativistic fire hose
101: instability).
102: \end{abstract}
103:
104: \keywords{ISM: individual (Crab Nebula) ---
105: ISM: jets and outflows ---
106: pulsars: individual (Crab Pulsar) ---
107: stars: neutron ---
108: supernova remnants}
109:
110: %% From the front matter, we move on to the body of the paper.
111: %% In the first two sections, notice the use of the natbib \citep
112: %% and \citet commands to identify citations. The citations are
113: %% tied to the reference list via symbolic KEYs. The KEY corresponds
114: %% to the KEY in the \bibitem in the reference list below. We have
115: %% chosen the first three characters of the first author's name plus
116: %% the last two numeral of the year of publication as our KEY for
117: %% each reference.
118:
119: \newpage
120:
121: \section{Introduction
122: \label{sec:gem1}}
123: Many young pulsars in supernova remnants
124: are embedded in synchrotron nebulae,
125: known as pulsar wind nebulae (PWN),
126: containing relativistic
127: electrons and magnetic flux emitted by the central object.
128: Multiwavelength imaging at subarcsecond resolution
129: reveals that many PWN share a common morphology,
130: consisting of
131: (i) bipolar jets of unequal brightness, directed along
132: the (inferred) pulsar spin axis and terminated by
133: one or more bright knots,
134: and (ii) fibrous arcs, or `wisps', that are cylindrically symmetric
135: about the spin axis, concave or convex with respect to
136: the pulsar, and associated with an X-ray ring and torus.
137: Objects known to display this morphology,
138: likened by some authors to the shape of a crossbow,
139: include the Crab \citep{hes95,wei00,bie01,hes02,sol03},
140: Vela \citep{pav01},
141: G320.4$-$1.2 \citep{gae02},
142: G54.1$+$0.3 \citep{lu02},
143: 3C58 \citep{sla02},
144: SNR 0540$-$69.3 \citep{got00},
145: and G0.9$+$0.1 \citep{gae01},
146: imaged variously by the
147: {\em Chandra X-Ray Observatory (CXO)},
148: {\em Hubble Space Telescope (HST)},
149: Very Large Telescope (VLT),
150: and Very Large Array (VLA),
151: although there are counter-examples too,
152: e.g.\ G11.2$-$0.3 \citep{kas01}.
153: The knots and wisps are interpreted, respectively,
154: as the polar and equatorial zones
155: of the magnetized, collisionless shock
156: terminating the pulsar wind
157: \citep{gal94,hes95,gae02,kom03,spi04}.
158: The brightness asymmetries are ascribed to Doppler boosting.
159: Ideas for explaining jet collimation
160: and the observed jet-torus structure
161: include magnetic hoop stress in the relativistic wind
162: upstream from the shock
163: \citep{beg94,bog01},
164: the anisotropic energy flux in a force-free, monopole wind
165: \citep{lyu01,lyu02,kom03,del04},
166: magnetic hoop stress or hydromagnetic instabilities
167: in the subsonic, downstream flow
168: \citep{beg92,beg98,lyu01,lyu02,mel02,mel04},
169: interactions with moving ejecta
170: \citep{pav03},
171: and the helical structure of a
172: wave-like, displacement-current-dominated wind
173: \citep{uso94,hes95,mel96,mel98,mel02}.
174:
175: The subarcsecond features of PWN are highly variable.
176: In the Crab PWN,
177: the optical wisps are seen to change brightness and position
178: in time-lapse {\em HST} images every six days,
179: receding from the pulsar concentrically in a wave-like pattern
180: at $\approx 0.5c$,
181: while the optical knots jump around erratically
182: on the same time-scale
183: \citep{hes95,hes02}.
184: In contemporaneous {\em CXO} images, taken every 22 days,
185: the X-ray ring inside the torus is resolved into
186: more than 20 knots that brighten and fade irregularly,
187: while the X-ray jet transports blobs of material and
188: bow-wave-shaped structures radially outward at $\approx 0.4c$
189: \citep{hes02,mor02}.
190: In time-lapse {\em CXO} images of the Vela PWN,
191: \citet{pav01,pav03} discovered changes of up to 30 per cent
192: over several months
193: in the brightnesses and spectra of the wisps, jet and
194: centrifugal ($0.3$--$0.6c$) knots,
195: coherent bending of the outer jet over 16 days,
196: and changes in knot brightness over just two days.
197: The variability has been ascribed to an
198: ion-cyclotron instability at the shock front \citep{spi04},
199: to a synchrotron cooling instability \citep{hes02},
200: and to a nonlinear Kelvin-Helmholtz instability
201: \citep{beg99}.
202:
203: In this paper, we report on the first near-infrared,
204: adaptive-optics observations of the wisps and jet
205: of the Crab PWN.
206: The data offer high angular resolution
207: (optimum $0.\arcsec19$, average $0.\arcsec23$ in $K'$),
208: high time resolution
209: ($0.6\,{\rm ks}$, resolving the light-crossing and
210: ion-cyclotron time-scales of the narrowest features for
211: the first time),
212: and the first feature-specific continuum color
213: spectra extending
214: from near-infrared to ultraviolet wavelengths.
215: The observations are described in \S\ref{sec:gem2},
216: the light curves and spectra of individual features
217: are presented in
218: \S\ref{sec:gem3} and \S\ref{sec:gem4},
219: and the results are interpreted physically in \S\ref{sec:gem5}.
220:
221: \section{Observations
222: \label{sec:gem2}}
223: The center of the Crab PWN was observed on
224: 2002 February 6, 7, and 8
225: with the Hokupa'a/QUIRC adaptive optics instrument
226: on the Gemini North Telescope
227: (proposal GN-2002A-Q-16).
228: Hokupa'a is a natural guide star, curvature sensing system
229: with 36 elements, coupled to a near-infrared camera,
230: QUIRC, consisting of a $1024\times 1024$ HgCdTe array
231: with plate scale $0.\arcsec0200$,
232: dark current $< 0.1$ electrons per second,
233: and read-out noise 15--30 electrons.
234: The system is described in detail by \citet{gra98}.
235: Hokupa'a is capable of locking on to
236: point sources with $R \lesssim 15\,{\rm mag}$ nominally
237: and $R \lesssim 17\,{\rm mag}$ in practice.
238: It is therefore ideal for imaging the Crab PWN,
239: where there are two suitable guide stars within
240: $4\arcsec$ of the center of the nebula
241: (cf.\ isoplanatic radius $\approx 30\arcsec$):
242: a field star (Star I) at
243: $\alpha(2000)=05^{\rm h}34^{\rm m}32.2^{\rm s}$ and
244: $\delta(2000)=+22^\circ00'57''$,
245: with $R=14.8\pm0.2\,{\rm mag}$,
246: which we chose to use,
247: and the Crab pulsar,
248: with $R=16.3\,{\rm mag}$ pulse-averaged \citep{eik97}.
249: The wavelengths and bandpasses of the QUIRC $J$ and $K'$ filters
250: are $1.25\,\mu{\rm m}/0.17\,\mu{\rm m}$
251: and $2.12\,\mu{\rm m}/0.41\,\mu{\rm m}$ respectively.
252: We achieved resolutions of
253: ${\rm FWHM} = 0.\arcsec19$--$0.\arcsec30$ in $K'$ and
254: ${\rm FWHM} = 0.\arcsec24$--$0.\arcsec37$ in $J$
255: in this sequence of observations,
256: where FWHM refers to the full-width half-maximum of
257: the point spread function (PSF).
258: The seeing fluctuated by up to $\pm0.05\arcsec$ in $K'$
259: and $\pm0.08\arcsec$ in $J$ over intervals of $0.6\,{\rm ks}$.
260:
261: We obtained a sequence of $20\times4\times 0.12\,{\rm ks}$
262: exposures spaced in
263: a four-point, $3\arcsec$ dither pattern,
264: which were subsequently combined into 20 frames,
265: as follows:
266: $5\times K'$ then $3\times J$ (2002 February 6),
267: $3\times J$ then $5\times K'$ (2002 February 7),
268: and $4\times K'$ (2002 February 8).
269: Conditions were excellent on 2002 February 6.
270: On the following nights,
271: observations were affected by light cloud and wind.
272: Data were gathered without interruption during all three nights,
273: yielding a sampling time of $0.6\,{\rm ks}$ between frames
274: (after read-out), with two brief exceptions:
275: a five-point dither was accidentally performed
276: on 2002 February 6,
277: and two six-point dithers were required on 2002 February 8,
278: when the telescope lost its guiding.
279: All features were observed with good signal-to-noise.
280: In a typical $0.12\,{\rm ks}$ exposure in $K'$,
281: we accumulated $\approx 4.9\times 10^2$ counts/pixel
282: for the pulsar and guide star,
283: $\approx 4.5\times 10^1$ counts/pixel
284: for the brightest extended feature
285: (the sprite; see \S\ref{sec:gem3}),
286: and $\approx 2.7\times 10^1$ counts/pixel
287: for the faintest extended feature
288: (the faint wisp; see \S\ref{sec:gem3}),
289: {\em after} subtracting the nebula background
290: ($2.9$--$5.1\times 10^3$ counts/pixel, or
291: $1.1$--$1.9\,{\rm Jy}\,{\rm arcsec^{-2}}$).
292:
293: We assembled a data reduction pipeline in
294: IRAF\footnote{Image Reduction and Analysis Facility,
295: Gemini package, v.\ 3.1}
296: to subtract
297: bias, dark, and sky frames and divide by flat fields
298: in the standard way.
299: The three bright point sources in the field were excised
300: from the sky frames with care, to avoid creating false
301: shadows by oversubtraction.
302: After trimming a border 20 pixels wide to remove faulty
303: edge pixels, we median combined each set of four dithered
304: images, rejecting one high,
305: then trimmed a border $3\arcsec$ wide to exclude the region
306: where the four dithered exposures do not overlap.
307:
308: An important issue with any adaptive optics observation
309: is the degree by which the PSF
310: changes across the field of view.
311: In this work, we can quantify the effect directly by
312: examining the three point sources in the field
313: (the guide star, the pulsar, and a field star
314: $6\arcsec$ south of the pulsar, labeled Star II),
315: which happen to be well separated.
316: We find that the PSF is
317: nearly axisymmetric at the southern field star,
318: elongated in an east-west direction at the pulsar,
319: and elongated in a northeast-southwest direction at the guide star,
320: with minor and major axes in the ratio $\approx 0.8$
321: (although the isophotes are not strictly elliptical).
322: This essentially precludes accurate photometry of the inner knot,
323: located $0.\arcsec6$ from the pulsar,
324: for the reasons set forth in \S\ref{sec:gem3b},
325: without prejudicing photometry of large, extended
326: features like the wisps.
327: Another issue is how the PSF varies as a function of time.
328: We do not detect any change in the shape of the PSF
329: at the locations of the three field stars
330: when comparing isophotes from successive exposures by eye.
331: However, there is indirect evidence that slight yet rapid
332: changes do occur;
333: we find that the total flux from the guide star
334: within an aperture of radius $3.1$ ${\rm FWHM}$
335: fluctuates by $\pm 6$ ($\pm 4$) per cent in $K'$ ($J$)
336: over $0.6\,{\rm ks}$ {\em after} background subtraction
337: (\S\ref{sec:gem3a}),
338: accompanied by fluctuations of $\pm 9$ ($\pm 4$)
339: per cent in the southern field star.
340: These changes feed into the measurement uncertainties
341: calculated in \S\ref{sec:gem3} and \S\ref{sec:gem4},
342: as the field stars serve as flux calibrators.
343:
344: \section{Kilosecond variability
345: \label{sec:gem3}}
346: Figure \ref{fig:gem1} shows the center of the Crab PWN
347: at $0.\arcsec19$ resolution (FWHM) in $K'$,
348: as it appeared on 2002 February 6.
349: Its crossbow-like morphology is evident.
350: The subarcsecond features in the termination shock of the
351: pulsar wind are cylindrically symmetric about
352: the projected rotation axis and proper motion
353: of the pulsar,
354: determined from {\em HST} astrometry \citep{car99}.
355: \begin{enumerate}
356: \item
357: The {\em wisps} are interpreted as shock structures in the
358: equatorial plane of the pulsar wind (latitude $\pm10\deg$),
359: in the neighborhood of the X-ray ring and torus
360: \citep{hes95,wei00,hes02,mor02,spi04,kom03,del04}.
361: The faint and bright wisps to the northwest of the pulsar,
362: labeled in Figure \ref{fig:gem1},
363: mark ion-driven magnetic compressions
364: (at the first and second ion turning points)
365: in the ion-cyclotron model of the shock
366: \citep{gal94,spi04};
367: they are probably analogous to the features
368: labeled 5 and E by \citet{gae02}
369: in another PWN, G320.4$-$1.2.
370: The position and brightness of these two wisps,
371: several less prominent wisps,
372: and their fibrous substructure are known to
373: change on a time-scale as short as six days in the
374: optical \citep{hes02} and 22 days in X-rays
375: \citep{hes02,mor02}.
376: \item
377: The {\em sprite} can be interpreted as a polar shock,
378: lying on the rotation axis (colatitude $\pm15\deg$)
379: at or near the base of the polar X-ray jet
380: \citep{wei00,hes02,mor02}.
381: It can also be interpreted as a mid-latitude arch shock
382: between the polar outflow and equatorial
383: backflow in a pressure-confined, split-monopole nebula,
384: situated at the tangent point of the line of sight
385: (and therefore Doppler boosted)
386: \citep{kom03,del04}.
387: Its shape, doughnut-like with a central rod (\S\ref{sec:gem3b}),
388: changes irregularly
389: on the same time-scale as the wisps;
390: in {\em CXO} images, the sprite appears to be the launching point
391: for blobs and
392: `bow waves' ejected along the X-ray jet
393: \citep{hes02,mor02},
394: although these may also be unstable motions in the
395: vicinity of the mid-latitude arch shock
396: \citep{kom03}.
397: \item
398: The {\em inner knot} is a barely resolved,
399: flattened (\S\ref{sec:gem3c}) structure partly obscured
400: in Figure \ref{fig:gem1} by the pulsar PSF.
401: It too can be interpreted as a polar feature,
402: lying $\approx 5$ times closer to the pulsar than the sprite.
403: If the sprite marks the polar termination shock,
404: the inner knot sits in the unshocked pulsar wind,
405: and its physical origin is unknown
406: \citep{hes95,mel98,mel02}.
407: If the sprite is a mid-latitude arch shock,
408: the inner knot originates from a part of the arch shock
409: nearer the base of the polar jet,
410: which is pushed inward (relative to the wisps) because the energy flux
411: in a split-monopole wind is lower at the poles than at the equator
412: \citep{kom03}.
413: \item
414: A conical {\em halo} is visible at intermediate latitudes,
415: midway between the pulsar and faint wisp
416: in Figure \ref{fig:gem1},
417: the near-infrared counterpart of an optical feature
418: noted in {\em HST} data by \citet{hes95}.
419: We do not discuss the halo further in this paper,
420: as it is too faint for accurate near-infrared photometry.
421: \end{enumerate}
422:
423: In this section, we examine the variability of four features ---
424: the bright wisp, faint wisp, sprite, and inner knot ---
425: in the near-infrared
426: over time-scales as short as $0.6\,{\rm ks}$,
427: extending previous studies of the Crab PWN with
428: {\em HST} (sampling time six days)
429: and {\em CXO} (sampling time 22 days)
430: \citep{hes02,mor02}.
431: The first (and most challenging) step,
432: subtracting the time-dependent nebula background,
433: is discussed in \S\ref{sec:gem3a}.
434: Light curves of the features are presented in
435: \S\ref{sec:gem3b} and \S\ref{sec:gem3c}.
436:
437: \subsection{Nebula background
438: \label{sec:gem3a}}
439: It is difficult to characterize and hence subtract
440: the background in Figure \ref{fig:gem1},
441: because there is no unique way to disentangle
442: the contributions from the nebula and sky,
443: given that the nebula is ubiquitous, nonuniform,
444: and time-dependent.
445: The surface brightness $I_{\rm o}$ observed
446: in any pixel is the sum of flux from
447: the nebula background ($I_{\rm n}$)
448: and any feature ($I_{\rm f}$) occupying that pixel,
449: with
450: $I_{\rm o} =
451: (I_{\rm n} + I_{\rm f}) \eta_{\rm a} + I_{\rm a}$,
452: where $\eta_{\rm a}$ and $I_{\rm a}$ denote the
453: atmospheric absorption coefficient
454: and sky brightness respectively.
455: There are two problems in extracting $I_{\rm f}$
456: from $I_{\rm o}$.
457: First,
458: there are no pixels empty of both features and nebula,
459: so we cannot measure $I_{\rm a}$ directly.
460: In our analysis, we assume that
461: $\eta_{\rm a}$ and $I_{\rm a}$
462: are uniform across the field of view,
463: but both parameters vary markedly from exposure
464: to exposure, as quantified below.
465: Second,
466: $I_{\rm n}$ fluctuates stochastically from pixel to pixel,
467: so we cannot estimate $I_{\rm n}$ behind a feature
468: by interpolating $I_{\rm n}$ directly
469: from neighboring, feature-free pixels.
470:
471: To overcome these problems,
472: we average the observed brightness of all
473: feature-free pixels ($I_{\rm o}'$) in the field to obtain
474: $I_{\rm n} \eta_{\rm a} + I_{\rm a}
475: = \langle I_{\rm o}' \rangle$
476: and hence, approximately,
477: $I_{\rm f} \eta_{\rm a} =
478: I_{\rm o} - \langle I_{\rm o}' \rangle$
479: along any line of sight with a feature,
480: under the assumption (justified below) that there is no
481: large-scale gradient of nebula brightness across the field.
482: The uncertainty in this estimate of $I_{\rm f} \eta_{\rm a}$
483: is given by the width of the $I_{\rm o}'$ distribution,
484: measured below.
485: Without extra, exposure-specific information,
486: it is impossible to determine
487: $\eta_{\rm a}$ and $I_{\rm f}$ independently.
488: However, we are interested here in the variability of
489: features rather than their absolute brightness.
490: Consequently, we can normalize the observed brightness
491: of any feature, $I_{\rm o,f}$,
492: to the observed flux $I_{\rm o,g}$
493: of an intrinsically steady point source $I_{\rm g}$
494: (e.g.\ the guide star or pulsar),
495: after subtracting the nebula background,
496: to obtain
497: $I_{\rm f}/I_{\rm g} =
498: (I_{\rm o,f} - \langle I_{\rm o}' \rangle ) /
499: (I_{\rm o,g} - \langle I_{\rm o}' \rangle ) $.
500: If it were necessary to determine $I_{\rm f}$ absolutely,
501: we would need to measure $I_{\rm g}$ independently,
502: e.g.\ from $J$ and $K'$ photometry of the
503: the Crab pulsar \citep{eik97}.
504:
505: How accurate is the above approach?
506: A histogram of pixel counts for a single $K'$ frame,
507: excluding point sources and extended features,
508: is presented in Figure \ref{fig:gem2}$a$
509: (solid curve).
510: The mean, median, and standard deviation ($\delta I$)
511: of the distribution in Figure \ref{fig:gem2}$a$
512: are 5033, 5024, and 13.2 counts respectively.
513: The distribution is not Gaussian,
514: cutting off sharply at $\pm 3.5 \delta I$,
515: and it is narrower than Poisson
516: ($\delta I =
517: 0.28 \langle I_{\rm o}' \rangle^{1/2}$)
518: because $I_{\rm n}$ is correlated in neighboring pixels.
519: For our $20$ images,
520: we find $\delta I$ in the range $10$--$30$ counts,
521: while the median varies markedly in the range
522: $2.9$--$5.1$ kcounts.
523: These statistics are corroborated by the dashed curve
524: in Figure \ref{fig:gem2}$a$,
525: a histogram of pixel counts averaged over
526: $50\times 50$ pixel blocks
527: (chosen to roughly match the dimensions of
528: the knot-like features of interest).
529: By inspecting the image directly,
530: we isolate the blocks that appear to be empty of features,
531: obtaining a mean, median and standard deviation of
532: $5033$, $5024$, and $13.4$
533: counts per block, in close agreement with the solid curve
534: (they are indistinguishable to the eye).
535: We also verify by inspection that there is no large-scale
536: gradient in counts per block across the field of view,
537: confirming that $\eta_{\rm a}$ and $I_{\rm a}$ are uniform
538: within the statistical uncertainty $\delta I$.
539: Finally, in Figure \ref{fig:gem2}$b$, we present the
540: histogram of pixel counts in an annular aperture
541: of radius $6.0$ ${\rm FWHM}$ centered on the guide star.
542: Annular and polygonal apertures are provided for
543: background measurements in the Gemini IRAF software
544: and are employed in \S\ref{sec:gem3b} and \S\ref{sec:gem3c}.
545: The statistics are consistent with
546: Figure \ref{fig:gem2}$a$.
547: We find that the median background in the annulus
548: differs by at most $8$ counts from the median of the field
549: in all $20$ images,
550: well within the standard deviation $\delta I$,
551: and is arguably a more accurate estimate of the
552: background locally.
553:
554: We estimate the uncertainty in our fluxes as follows.
555: The absolute uncertainty in $I_{\rm f} \eta_{\rm a}$,
556: the total flux minus the background,
557: is given by $(e_1^2 + e_2^2 + e_3^2)^{1/2}$,
558: where $e_1$ is the square root of the counts
559: after background subtraction
560: (corrected for the ADU-photon ratio),
561: $e_2$ equals $N_{\rm f}^{1/2} \delta I$
562: ($N_{\rm f}$ is the number of pixels in the aperture
563: enclosing the feature),
564: and $e_3$ is given by $(N_{\rm f}/N_{\rm s})^{1/2} e_2$
565: ($N_{\rm s}$ is the number of pixels in the aperture
566: estimating the sky).
567: A similar uncertainty attaches to the guide star
568: $I_{\rm g} \eta_{\rm a}$.
569: Note that $e_1$ represents the Poisson
570: fluctuation in the intrinsic flux of the feature;
571: $e_2$ measures the uncertainty
572: in the background contribution to the total flux,
573: characterized by Figure \ref{fig:gem2}$a$ and $\delta I$
574: (not Poissonian);
575: and $e_3$ is the uncertainty in the background level
576: subtracted from the total flux,
577: corrected for the relative sizes of the sky and
578: feature apertures.
579:
580: \subsection{Termination shock: wisps and sprite
581: \label{sec:gem3b}}
582: In this section,
583: we examine the variability of the
584: equatorial and polar zones
585: (wisps and sprite)
586: of the termination shock in the Crab PWN.
587:
588: Figure \ref{fig:gem3} shows two enlarged images
589: of the bright and faint wisps in $J$ band,
590: taken $1.2\,{\rm ks}$ apart on 2002 February 7,
591: after nebula subtraction and normalization
592: to the guide star
593: ($I_{\rm f}/I_{\rm g}$; see \S\ref{sec:gem3a}).
594: The frames should be identical if there is no change in
595: the intrinsic brightness of the wisps,
596: yet differences between them are readily apparent
597: (although the features are not displaced).
598: To quantify the changes, we use the {\tt polymark}
599: tool in IRAF to specify a polygonal aperture
600: enclosing each wisp, as drawn in Figure \ref{fig:gem3}.
601: The fluxes enclosed by the apertures,
602: after nebula subtraction and normalization,
603: are plotted as functions of time
604: in Figure \ref{fig:gem4}.
605: $J$- and $K'$-band data are both displayed;
606: uncertainties are calculated according to the recipe
607: in \S\ref{sec:gem3a}.
608: For the bright wisp,
609: we find maximum flux changes (relative to the mean level)
610: of $\pm 24\pm4$ per cent in $J$
611: and $\pm 12\pm7$ per cent in $K'$,
612: occurring in the space of $1.2\,{\rm ks}$ on
613: 2002 February 7,
614: and smaller fluctuations at other times.
615: Moreover, the light curves of the bright and faint wisps
616: appear correlated to some degree in both filters.
617: The detection of variability is marginal in $K'$
618: but more statistically significant in $J$,
619: where the nebula background is less prominent.
620: We find, by experimentation, that the results are
621: essentially independent of the choice of aperture,
622: while the time-dependent PSF has a minimal effect
623: on the measured flux of the extended features
624: (for the guide star,
625: the effect is included in the measurement uncertainty;
626: see \S\ref{sec:gem2}).
627: Nevertheless, new observations ---
628: preferably by an independent party
629: using a different instrument ---
630: need to be made
631: before variability of the wisps on such short time-scales
632: can confidently be claimed.
633:
634: Figure \ref{fig:gem5} shows two enlarged images of
635: the sprite in $J$ band,
636: taken on 2002 February 6 and 8.
637: It is interesting to note its doughnut-like structure,
638: symmetric about the pulsar's rotation axis,
639: as well as the short, bent rod emerging from its center,
640: seen clearly here for the first time and corroborating
641: the observation by \citet{hes02} that the sprite is often
642: center-filled (especially in X-rays).
643: Knot-like features in another PWN, G320.4$-$1.2,
644: numbered 2 and 3 by \citet{gae02},
645: may be analogs of the sprite.
646:
647: In common with the wisps,
648: there are visible differences in the brightness
649: (but not the position) of the sprite in the two images,
650: even after nebula subtraction and normalization ---
651: not just the absolute brightness, but also,
652: more significantly, the brightness contrast between
653: the rod and doughnut,
654: which is less likely to be affected by PSF nonuniformity
655: and imperfect sky/nebula subtraction.
656: (The flux of the guide star is
657: equal to within 4 per cent in the two snapshots.)
658: To quantify the brightness changes, we define apertures
659: enclosing the rod and the whole sprite,
660: and plot the nebula-subtracted, normalized aperture flux
661: $I_{\rm f}/I_{\rm g}$
662: versus time in Figure \ref{fig:gem6}.
663: We measure the following maximum flux changes
664: (relative to the mean):
665: $\pm13\pm4$ per cent (rod, $J$),
666: $\pm15\pm7$ per cent (rod, $K'$),
667: $\pm9\pm4$ per cent (whole sprite, $J$),
668: and
669: $\pm8\pm7$ per cent (whole sprite, $K'$).
670: As for the wisps,
671: the detection of variability is marginal in $K'$
672: and somewhat more significant in $J$,
673: especially for the rod,
674: as seen in Figure \ref{fig:gem5}.
675:
676: \subsection{Pulsar wind: inner knot
677: \label{sec:gem3c}}
678: The inner knot,
679: discovered by \citet{hes95} in optical {\em HST} data,
680: is displaced $0.\arcsec65$ from the pulsar
681: along the axis of symmetry of the PWN,
682: and is resolved by {\em HST} to be
683: $\approx 0.\arcsec2$ thick \citep{hes95}.
684: No counterpart has been detected unambiguously
685: at X-ray wavelengths,
686: although there is a hint of a southeasterly `bump'
687: protruding from the pulsar in {\em CXO} images,
688: e.g.\ in Figure 5 of \citet{hes02}.
689: It is also possible that an analogous feature,
690: named feature 1,
691: has been discovered in {\em CXO} images of
692: another PWN, G320.4$-$1.2 \citep{gae02}.
693: In Figure \ref{fig:gem7}, we present
694: a brightness map of the inner knot,
695: after subtraction of the PSF.
696: It is clear, from the isophotes (solid contours) in particular,
697: that the feature is flattened, not spherical,
698: although it is hard to discern its shape exactly
699: because it is barely resolved in our
700: highest-resolution $K'$ data
701: and the PSF subtraction is imperfect.
702:
703: Photometry of the inner knot is complicated by its proximity
704: to the pulsar,
705: whose flux contaminates the knot unpredictably
706: from image to image as the seeing fluctuates.
707: The standard approach, modeling and subtracting the pulsar PSF,
708: is attempted in Figure \ref{fig:gem7},
709: but the result is unreliable; see \S\ref{sec:gem2}.
710: Faced with these difficulties,
711: we test for variability of the inner knot
712: without measuring its brightness directly
713: in two complementary ways.
714: In the first test, we measure the fluxes
715: $I_{\rm p}(r_1)$ and $I_{\rm p}(r_2)$
716: (after subtracting $\langle I_{\rm o}' \rangle$)
717: enclosed by two circular apertures centered on the pulsar,
718: of radii $r_1=0.75$ ${\rm FWHM}$
719: (including as much PSF as possible but
720: excluding most of the knot)
721: and $r_2=5.0$ ${\rm FWHM}$
722: (including the PSF and knot).
723: The apertures are drawn in Figure \ref{fig:gem7}.
724: The ratio of these fluxes,
725: after nebula subtraction, would be the same
726: as the ratio of the fluxes
727: $I_{\rm g}(r_1)$ and $I_{\rm g}(r_2)$
728: enclosed by identical apertures around the guide star
729: {\em if there were no inner knot},
730: because the cylindrically averaged PSF is uniform
731: to within $\pm6$ ($\pm4$) per cent in $K'$ ($J$)
732: across the field of view (see \S\ref{sec:gem2}).
733: Therefore, the flux difference
734: $I_{\rm k}/I_{\rm g}=
735: I_{\rm p}(r_2)/I_{\rm g}(r_2) -
736: I_{\rm p}(r_1)/I_{\rm g}(r_1)$
737: can be attributed to the presence of the inner knot,
738: and any change in $I_{\rm k}/I_{\rm g}$
739: from image to image is evidence that the
740: inner knot varies intrinsically.
741: In Figure \ref{fig:gem8}, we plot
742: $I_{\rm k}/I_{\rm g}$ as a function of time for
743: the full sequence of observations.
744: The data are consistent with no variability,
745: within the measurement uncertainties.
746: For example, in the $K'$ band, we find
747: maximum peak-to-peak changes in $I_{\rm k}/I_{\rm g}$
748: of $0.045\pm0.060$ over $1.2\,{\rm ks}$
749: and
750: $0.054\pm0.063$ over $48$ hours.
751:
752: A second test provides a consistency check:
753: we measure directly the flux enclosed by a circular
754: aperture of radius 10 pixels, centered on the knot.
755: The flux thus measured includes leakage from the pulsar PSF,
756: the amount of which varies from
757: image to image along with the seeing,
758: as noted above,
759: but strong intrinsic variations
760: in knot brightness could still overwhelm this effect.
761: In Figure \ref{fig:gem9}, we plot the flux in the
762: 10-pixel aperture as a function of time,
763: after nebula subtraction and guide star normalization.
764: The result is consistent with Figure \ref{fig:gem8}:
765: there is no significant detection of variability
766: within the measurement uncertainties,
767: with a maximum peak-to-peak change
768: of $0.011\pm0.007$ over $1.2\,{\rm ks}$
769: on 2002 February 7.
770:
771: On the strength of the data presented here,
772: we are unable to say whether or not the inner knot
773: is variable on time-scales of $0.6\,{\rm ks}$
774: to $48$ hours. New observations ---
775: preferably by an independent party using a
776: different instrument ---
777: are required to settle the issue,
778: and dedicated PSF calibration frames will be
779: essential if adaptive optics are used.
780:
781: \section{Feature-specific color spectra
782: \label{sec:gem4}}
783:
784: \subsection{Near infrared: wisps, sprite, and inner knot
785: \label{sec:gem4a}}
786: In this section, we measure the $J$-to-$K'$ color spectra
787: of the faint wisp, bright wisp, sprite, rod, and inner knot.
788: As these features may vary on time-scales shorter than
789: the minimum interval between exposures,
790: we sum the 14 $K'$ images and 6 $J$ images in our data set
791: to obtain time-averaged spectra.
792:
793: To the best of our knowledge,
794: calibrated $J$-to-$K'$ spectra of Stars I and II have not been published.
795: Therefore, to calibrate the spectra of the various
796: subarcsecond features in the PWN,
797: we are forced to redo the photometry in
798: \S\ref{sec:gem3} by normalizing nebula-subtracted fluxes
799: with respect to the pulsar,
800: whose phase-averaged, near-infrared, color spectrum
801: was determined by \citet{eik97}
802: using the Solid State Photomultiplier
803: on the Multiple Mirror Telescope.
804: After dereddening, the pulsar's spectrum is fairly flat
805: in the near infrared ($F_\nu \propto \nu^{0.36}$);
806: see the third row of Table \ref{tab:gem1}.
807: Note that the fluxes in \citet{eik97} include the inner knot
808: (unresolved from the ground).
809: By comparing with the total flux of the pulsar plus knot in our images,
810: we derive a zero magnitude reference for calibrated
811: aperture photometry of the extended features (e.g.\ wisps, sprite, and rod).
812:
813: To measure the flux of the knot, we remove the contribution from
814: the pulsar by scaling it to the azimuthally averaged PSFs
815: of Stars I and II,
816: such that the normalization is the same at a radius of 10 pixels.
817: The combined profile of the pulsar and knot is found to match
818: the stellar PSFs up to a radius of $\approx 20$ pixels
819: (to within $0.05\,{\rm mag}$ in $K'$ and $0.03\,{\rm mag}$ in $J$)
820: but deviates beyond due to the knot excess.
821: We measure the flux difference between radii of 20 and 60 pixels
822: (where the profile merges into the background)
823: and adjust the results in \citet{eik97} to give
824: the calibrated knot flux.
825: Note that the differential method of detecting the knot
826: ($I_{\rm k}/I_{\rm g}$)
827: employed in \S\ref{sec:gem3c} does not yield a calibrated flux.
828:
829: Time-averaged $J$- and $K'$-band fluxes are presented
830: in Table \ref{tab:gem1}
831: for the point-like and extended subarcsecond features identified
832: in Figure \ref{fig:gem1},
833: together with the spectral index $\alpha$ of each feature
834: assuming its flux density scales as $F_\nu \propto \nu^{\alpha}$
835: (as for synchrotron radiation,
836: although not necessarily for all nonthermal processes,
837: e.g.\ synchro-Compton radiation; see \S\ref{sec:gem5}).
838: The fluxes for the extended features are quoted per unit length,
839: although it is possible that we are marginally resolving the wisps
840: across their width.
841: The field stars are included for reference,
842: as their spectra have not been published previously.
843: Three sources of uncertainty,
844: summed in quadrature, contribute in Table \ref{tab:gem1}.
845: First,
846: the central wavelength of the $K$ filter used
847: by \citet{eik97}
848: is $0.08\,\mu{\rm m}$ greater than for the
849: QUIRC $K'$ filter.
850: Second, there is scatter in the 20-pixel flux and 20-to-60-pixel offset.
851: Third,
852: the extinction corrections are uncertain
853: ($0.63\pm 0.03\,{\rm mag}$ in $J$ and
854: $0.81\pm 0.03\,{\rm mag}$ in $K'$).
855:
856: Several interesting conclusions emerge from these data.
857: First, the inner knot is clearly the reddest feature in the region,
858: with $\alpha\approx {-0.8}$.
859: This is understandable if the inner knot is produced in the
860: pulsar wind by a different radiation mechanism than other features.
861: However, it is surprising if the inner knot is part of the
862: same arch shock that produces the sprite.
863: Second, the polar or mid-latitude sprite and rod have
864: flatter spectra ($\alpha={-0.21\pm0.13}$)
865: than the equatorial wisps ($\alpha={-0.56\pm0.12}$).
866: Yet all these features are synchrotron emitting elements
867: of the same termination shock, albeit at different latitudes.
868: Third, there is no feature in the region whose spectrum
869: matches smoothly from the near infrared to X-rays.
870: The theoretical implications of these results
871: are considered further in \S\ref{sec:gem5}.
872:
873: \subsection{Ultraviolet: inner knot
874: \label{sec:gem4b}}
875: We extend the spectrum of the inner knot
876: in Table \ref{tab:gem1} with an independent measurement
877: of the ultraviolet flux of this feature.
878: \citet{gul98} observed the Crab pulsar with the {\em HST}
879: Space Telescope Imaging Spectrograph (STIS) NUV-MAMA detector
880: on 1997 August 7 through the low dispersion G230L grating.
881: The observations were made using a $2\arcsec \times 2\arcsec$
882: aperture which included the inner knot,
883: with part of the exposure ($2 \times 2.4\,{\rm ks}$) in TIME-TAG mode.
884: We acquired the archival data, barycentered the photons using
885: standard STIS routines, and extracted a `slit' image using photons
886: from $0.25\,{\rm rad}$ of phase spanning the pulse minimum,
887: thereby gating out the pulsar.
888: Approximately two per cent of the unpulsed flux remained,
889: but, after subtracting a scaled version of the on-pulse PSF,
890: a clear excess was found at the projected position of the
891: inner knot $\approx 0.3\pm0.2\arcsec$ on one side of the pulsar.
892: The intensity profile agrees with that expected from a
893: one-dimensional collapse of direct HST images.
894: Assuming a flat $\alpha \approx 0$ spectral index over the NUV band
895: ($160$--$320\,{\rm nm}$),
896: we find a summed inner knot flux
897: $(3.3 \pm 0.1) \times 10^{-2}$ times the unpulsed flux of the pulsar.
898: De-reddening with the best fit value $E(B-V)=0.52$ \citep{sol00}
899: yields
900: $F_\nu = 0.12 \pm 0.01\,{\rm mJy}$
901: for wavelengths in the range
902: $0.16$--$0.32\,\mu{\rm m}$.
903: Note that the ultraviolet flux is consistent with the
904: near-infrared spectrum measured in \S\ref{sec:gem4a},
905: which extrapolates to give $F_\nu = 0.11\,{\rm mJy}$ at
906: $0.32\,\mu{\rm m}$ for $\alpha=-0.8$.
907:
908: \section{Discussion
909: \label{sec:gem5}}
910: In this paper, we report on the first near-infrared,
911: adaptive-optics observations of the wisps and jet of the
912: Crab PWN, comprising 20 $J$- and $K'$-band snapshots
913: taken at $0.\arcsec19$--$0.\arcsec37$ resolution
914: with the Hokupa'a/QUIRC camera on the Gemini North Telescope.
915: The data contain tantalizing --- albeit inconclusive ---
916: evidence that subarcsecond features in the termination shock
917: of the Crab PWN vary intrinsically in $J$-band brightness
918: by $\pm24\pm4$ (wisps) and $\pm14\pm4$ (sprite) per cent
919: on time-scales as short as $1.2\,{\rm ks}$.
920: The principal sources of uncertainty
921: are the nonuniform, unsteady nebula background and PSF.
922: The data also suggest that the near-infrared spectra
923: of polar features in the termination shock are flatter
924: (e.g.\ sprite, $F_\nu \propto \nu^{-0.21\pm0.13}$)
925: than the spectra of the equatorial wisps
926: ($F_\nu \propto \nu^{-0.56\pm0.12}$),
927: except for the steep-spectrum inner knot
928: ($F_\nu \propto \nu^{-0.8}$),
929: which may lie in the unshocked pulsar wind.
930: This result is supported by an independent measurement of the
931: ultraviolet flux of the inner knot,
932: obtained by reanalyzing archival, time-tagged, {\em HST} STIS data.
933:
934: \subsection{Ion cyclotron and fire hose instabilities
935: \label{sec:gem5a}}
936: Why, physically, might the nebula vary on time-scales
937: as short as $1.2\,{\rm ks}$?
938: According to one hypothesis,
939: modeled numerically by \citet{spi04},
940: the wind contains ions which
941: drive a relativistic cyclotron instability at the termination shock.
942: The instability exhibits limit cycle dynamics:
943: ion bunches and
944: magnetic compressions are launched downstream periodically
945: at roughly half the ion-cyclotron period,
946: $\frac{1}{2} T_{\rm i} =
947: \pi A_{\rm i} m_{\rm p}\gamma_{\rm i} /
948: Z_{\rm i} e B$,
949: where $A_i$ and $Z_i$ are the
950: atomic number and charge, $\gamma_i$ is the preshock
951: Lorentz factor, and $B$ is the postshock magnetic field.
952: In order to fit the separation of the innermost wisps,
953: one must take $T_{\rm i} \approx 3.2\times 10^7\,{\rm s}$,
954: consistent with $\gamma_{\rm i}=7.8\times 10^5$ and
955: $B=16\,\mu{\rm G}$ at a radial distance $r=0.1\,{\rm pc}$
956: from the pulsar (for $A_{\rm i} / Z_{\rm i}=1$)
957: \citep{spi04}.
958: Faster variability is expected nearer the pulsar,
959: because the magnetic field in the wind scales
960: as $B=16(r/0.1\,{\rm pc})^{-1}\,\mu{\rm G}$.
961: \footnote{Faster variability is also expected at certain
962: special phases in the month-long ion cycle,
963: e.g. in the neighborhood of a moving wisp,
964: where the plasma is stirred up
965: (A. Spitkovsky, private communication).}
966: However, we find
967: $T_{\rm i}\approx 2\times 10^7\,{\rm s}$
968: and
969: $T_{\rm i}\approx 2\times 10^6\,{\rm s}$
970: at the sprite ($r\approx 0.05\,{\rm pc}$)
971: and inner knot ($r\approx 0.007\,{\rm pc}$) respectively,
972: slower than the variability observed.
973: Relativistic Doppler boosting does not improve the agreement;
974: the time-scale $\propto \gamma_{\rm i}^{-2}$
975: is unchanged downstream
976: ($\gamma_{\rm i}\sim 1$) and too short upstream
977: ($\gamma_{\rm i}\sim 10^6$).
978: The electron-cyclotron period,
979: $T_{\rm e}=(m_{\rm e}Z_{\rm i}/m_{\rm p}A_{\rm i})T_{\rm i}$,
980: does fall in the observed range,
981: but it is hard to see how to maintain
982: coherent limit-cycle dynamics in the electrons when they are
983: randomized rapidly at the shock by ion-driven magnetosonic waves
984: \citep{gal94,spi04}.
985: We are therefore inclined to rule out a cyclotron origin
986: of the observed kilosecond variability in the near infrared.
987:
988: The argument against a cyclotron origin of the kilosecond
989: variability assumes that all the energy in the unstable
990: (compressional) ion-cyclotron-magnetosonic waves
991: resides in the fundamental. This need not be so.
992: The frequency spectrum of the waves is quite flat in
993: one-dimensional simulations \citep{hos92};
994: significant power is deposited at high harmonics
995: (up to orders $\sim m_{\rm p}/m_{\rm e}$)
996: provided that parametric three-wave decays do not destroy
997: the coherence of the waves, a plausible concern in a realistic,
998: three-dimensional plasma (J. Arons, private communication).
999:
1000: Another possible mechanism,
1001: applicable especially to the knots in the polar jet,
1002: is the relativistic fire hose instability,
1003: driven by anisotropy of the kinetic pressure
1004: parallel ($P_\parallel$)
1005: and perpendicular ($P_\perp$) to the magnetic field $B$.
1006: \citet{noe69} showed that the growth time $T_{\rm fh}'$
1007: in the bulk frame of the jet
1008: (denoted by primes) is given by
1009: \begin{equation}
1010: \frac{1}{T_{\rm fh}'} =
1011: \frac{eB'}{m_{\rm e} \langle\gamma'\rangle}
1012: \frac{1.3-B'^2/[\mu_0 (P'_\parallel-P'_\perp)]}
1013: {(1+c^2/v_{\rm A}'^2)^{1/2} + 8}~,
1014: \label{eq:gem1}
1015: \end{equation}
1016: where $\langle\gamma'\rangle$ is the thermal Lorentz factor,
1017: $v_{\rm A}'^2=B'^2/\mu_0\rho'$ is the Alfv\'{e}n speed,
1018: and $\rho'$ is the density of the cold ion background.
1019: As long as the condition
1020: $P_\parallel'-P_\perp'>0.77B'^2/\mu_0$
1021: is met, the minimum growth time is given by
1022: $T_{\rm fh,min}'=0.14 m_{\rm e} \langle\gamma'\rangle/eB'$
1023: for $v_{\rm A}'\gg c$.
1024: Upon Lorentz transforming to the observer's frame,
1025: we obtain
1026: (i)
1027: $T_{\rm fh,min}=
1028: 0.14 m_{\rm e} \langle\gamma\rangle/eB$
1029: if $B$ is radial in the jet,
1030: or (ii)
1031: $T_{\rm fh,min}=
1032: 0.14 \gamma_{\rm i} m_{\rm e} \langle\gamma\rangle/eB$
1033: if $B$ is helical or toroidal in the jet,
1034: assuming
1035: $\langle\gamma\rangle \approx
1036: \gamma_{\rm i}\langle\gamma'\rangle$.
1037: (The validity of the last assumption depends subtly on the
1038: exact form of the electron distribution.)
1039: Immediately upstream from the termination shock
1040: of the jet, we have $\langle\gamma\rangle = 1$,
1041: $B=2.7\,\mu{\rm G}$, and $\gamma_{\rm i}=7.8\times 10^5$,
1042: implying $T_{\rm fh,min}=2.3\,{\rm ks}$ in scenario (ii).
1043: Immediately downstream from the termination shock,
1044: we have $\langle\gamma\rangle \approx 1\times 10^6$,
1045: $B=8\,\mu{\rm G}$, and $\gamma_{\rm i}=1.1$,
1046: implying $T_{\rm fh,min}\approx 1.1\,{\rm ks}$
1047: in scenarios (i) and (ii) ---
1048: intriguingly close to the observed time-scale.
1049: Note that the fire hose instability only occurs
1050: for $P_\parallel' > P_\perp'$;
1051: in the reverse situation, a mirror instability exists for
1052: $P_\perp'-P_\parallel' > 8.9B'^2/\mu_0$,
1053: with growth time $\sim T_{\rm fh}'$.
1054: \citet{gal94} argued that $P_\parallel'/P_\perp'$
1055: increases from zero to unity downstream from the shock ---
1056: the adiabatic index
1057: $(3+P_\parallel'/P_\perp')/(2+P_\parallel'/P_\perp')$
1058: decreases from 3/2 to 4/3 as pitch-angle scattering
1059: isotropizes the electrons ---
1060: but $P_\parallel' > P_\perp'$ cannot be excluded.
1061:
1062: Despite appearances, it is unlikely that
1063: the fire hose instability causes the
1064: unsteady, serpentine motions observed in the Vela X-ray jet
1065: on time-scales between one day and several weeks
1066: \citep{pav03},
1067: because the growth time appears to be too short.
1068: We estimate
1069: $T_{\rm fh,min}=
1070: 0.14 \gamma_{\rm i} m_{\rm e} \langle\gamma\rangle/eB
1071: \approx 0.1 (\sigma/10^{-3})^{1/2}\,{\rm ks}$,
1072: taking
1073: $\langle\gamma\rangle=7\times 10^4$
1074: [pair multiplicity $\approx 10^3$; see Figure 17 of
1075: \citet{hib01}]
1076: and
1077: $B=0.15\sigma^{1/2}\,{\rm mG}$ (radial or toroidal),
1078: where $\sigma$ is the ratio of Poynting to kinetic energy flux
1079: at the base of the jet.
1080: \citet{pav03} suggests an alternative scenario,
1081: in which the end of the jet is bent by an external wind
1082: while the knots in the body of the jet are produced by
1083: hydromagnetic (kink and sausage) instabilities on the local
1084: Alfv\'{e}n time-scale.
1085:
1086: Our time-lapse observations resolve the light-crossing
1087: time-scale of the smallest features in the field,
1088: e.g.\ $2.2$ days for $0.\arcsec19$ at $2.0\,{\rm kpc}$.
1089: Therefore, if the kilosecond variability we observe is real,
1090: it must arise from
1091: (i) a pattern traveling at a superluminal phase speed,
1092: or (ii) relativistic Doppler boosting in the
1093: upstream collimated outflow
1094: ($\gamma_{\rm i}^{-2}$ times the light-crossing time-scale;
1095: cf.\ millisecond variability of unresolved gamma-ray bursters).
1096: Our results are consistent with previous observations that
1097: also detected significant variations on, or faster than,
1098: the light-crossing time-scale:
1099: \citet{hes02} observed the optical knots in the Crab PWN
1100: to vary over six days,
1101: and \citet{pav03} observed the X-ray jet in the Vela PWN
1102: to vary over just two days.
1103:
1104: \subsection{Radiation mechanisms
1105: \label{sec:gem5b}}
1106: The near-infrared spectral indices displayed
1107: in Table \ref{tab:gem1} are curious in several respects.
1108: First, the inner knot has a steeper spectrum than
1109: every other feature in the region --- not just in our data,
1110: where the $J$-band flux is uncertain to $\pm 50$ per cent,
1111: but also in data obtained with the Infrared Spectrometer
1112: And Array Camera on the VLT
1113: in $0.\arcsec65$--$0.\arcsec88$ natural seeing \citep{sol03}.
1114: One explanation is that the inner knot is physically different
1115: from the other features:
1116: the wisps and sprite are part of the termination shock
1117: and emit synchrotron radiation,
1118: e.g.\ from ion-cyclotron-heated electrons
1119: \citep{gal94},
1120: whereas the inner knot lies upstream in the unshocked pulsar wind
1121: and emits synchro-Compton radiation,
1122: e.g.\ from electrons heated by magnetic reconnection
1123: \citep{cor90,lyu01a}
1124: or parametric instabilities
1125: \citep{mel96,mel98,mel02}
1126: in a wave-like wind.
1127: This explanation conflicts with recent simulations
1128: which suggest that the inner knot is synchrotron emission
1129: from an arch shock between the polar outflow and
1130: equatorial backflow in the nebula
1131: \citep{kom03}.
1132: Unfortunately, we cannot discriminate between these two possibilities
1133: spectrally, because both synchrotron and synchro-Compton radiation
1134: yield $\alpha=(1-p)/2$
1135: given a power-law electron distribution
1136: $N(\gamma) \propto \gamma^{-p}$
1137: \citep{bla72,leu82}.
1138: [Monoenergetic electrons in a large-amplitude wave
1139: emit an inverse Compton spectrum
1140: $F_\nu \propto \nu$ at frequencies below
1141: $0.1\gamma^2 a^3 \Omega \sim 10^{16}\,{\rm Hz}$,
1142: where $a=1$--10 is the wave nonlinearity parameter
1143: \citep{mel96,mel98}
1144: and $\Omega$ is the
1145: pulsar spin frequency, but this low-frequency tail is modified to
1146: $F_\nu \propto \nu^{(1-p)/2}$ for power-law electrons.]
1147: Instead, we propose that the near-infrared polarizations of the
1148: inner knot and the sprite be measured.
1149: Theory predicts that synchrotron and synchro-Compton
1150: radiation are polarized perpendicular and parallel
1151: to the magnetic field respectively;
1152: one expects 50--80 per cent linear polarization from a linearly
1153: polarized large-amplitude wave if $0.5\leq p\leq 5$
1154: \citep{bla72}.
1155: Therefore, assuming that the magnetic geometry of the wave-like wind
1156: is similar at the inner knot and sprite
1157: (e.g.\ helical at high latitude),
1158: the polarization vectors of the two features are expected to be
1159: perpendicular if the inner knot originates from the unshocked
1160: pulsar wind and parallel if it is an arch shock.
1161: Moreover, if the inner knot originates from
1162: the unshocked wind, it should also be circularly polarized
1163: (degree $\sim a^{-1}$, independent of $\nu$).
1164:
1165: A second curious property of the spectra in Table \ref{tab:gem1}
1166: is that the bright and faint wisps are steeper than the
1167: sprite and rod. In this case, there is no doubt that both sets
1168: of features are shock-related synchrotron emission
1169: at equatorial and polar latitudes respectively,
1170: whether the sprite is located
1171: at the working surface of the polar jet \citep{hes02}
1172: or along the arch shock \citep{kom03,del04}.
1173: So why do the spectra differ?
1174: One possibility is that the polarization of the large-amplitude wave
1175: in the wind zone,
1176: which is linear at the equator and circular at the pole,
1177: affects the acceleration physics in the shock ponderomotively;
1178: tentative indications to this end are emerging from recent
1179: particle-in-cell simulations
1180: (O.\ Skjaeraasen, private communication).
1181:
1182: Table \ref{tab:gem1} raises a third puzzle:
1183: the near-infrared spectra of some features do not extrapolate smoothly
1184: to optical and X-ray wavelengths, while others do.
1185: For example, the {\em HST} $V$-band surface brightness of the bright wisp
1186: and the $V$-band flux of the sprite
1187: are measured to be $49\,\mu{\rm Jy\,arcsec^{-1}}$ and
1188: $4.1\,\mu{\rm Jy}$ respectively
1189: \citep{hes95};
1190: the same quantities, extrapolated from Table \ref{tab:gem1},
1191: are predicted to be
1192: $\approx 42\,\mu{\rm Jy\,arcsec^{-1}}$ and
1193: $\approx 15\,\mu{\rm Jy}$ respectively.
1194: VLT and {\em HST} observations show that
1195: the spectrum of the inner knot extends smoothly from near-infrared to
1196: optical wavelengths, with $\alpha=-0.75\pm0.15$
1197: \citep{sol03},
1198: and this is corroborated (within larger uncertainties) by our data.
1199: On the other hand, extrapolation of the VLT spectrum of the inner knot
1200: to $0.16$--$0.32\,\mu{\rm m}$ yields
1201: $F_\nu = 0.35\pm 0.05\,{\rm mJy}$,
1202: significantly greater than the ultraviolet flux measured in \S\ref{sec:gem4b}.
1203:
1204: The spectrum of the bright and faint wisps
1205: is significantly steeper in X-rays than in the near infrared,
1206: with $\alpha_{\rm X} \approx -1.3$ \citep{wei00}
1207: and $\alpha_{\rm NIR} - \alpha_{\rm X} \approx 0.7$,
1208: comparable to the steepening expected from synchrotron cooling
1209: (Y.\ Lyubarsky, private communication).
1210: The synchrotron cooling time for near-infrared-emitting electrons,
1211: $t_{\rm cool}=2.8\times 10^4 (B/16\,\mu{\rm G})^{-3/2}
1212: (\nu/2.4\times 10^{14}\,{\rm Hz})^{-1/2}\,{\rm yr}$,
1213: is much longer than the flow time $t_{\rm flow}$ across the wisps.
1214: On the other hand, the near-infrared spectral index of the sprite and rod
1215: ($\alpha = -0.21\pm 0.13$) is significantly shallower than the wisps
1216: and similar to the average {\em radio} spectral index of the nebula
1217: (Y.\ Lyubarsky, private communication).
1218: This is either coincidental or highly surprising.
1219: The sprite and rod are shock features in which one has
1220: $t_{\rm flow} \ll t_{\rm cool}$ for near-infrared-emitting electrons
1221: and kilosecond variability is observed;
1222: they reflect electron acceleration at the present time.
1223: The radio electrons
1224: reflect the history of electron acceleration over the lifetime of the nebula
1225: and should be unaffected by the present dynamics of the sprite.
1226: This paradox is encountered in a related context:
1227: \cite{bie01} observed that the radio and optical wisps travel radially
1228: in concert and display coordinated spectral index variations.
1229:
1230: We conclude by reiterating that the measurement uncertainties
1231: in our observations are substantial, due to the nonuniformity
1232: of the nebula and PSF. Consequently, the evidence for
1233: kilosecond variability and spectral differences,
1234: while tantalizing, is inconclusive.
1235: Improved observations --- preferably by an independent party
1236: using a different instrument --- are essential
1237: to clarify the situation.
1238:
1239: \acknowledgments
1240: We thank D. Barnes, A. Karick, M. O'Dowd, and J. Stevens
1241: for assisting with data reduction
1242: and A. Oshlack for advice regarding error estimation.
1243: This research was supported in part by
1244: Australian Research Council grant DP 0208735 (AM)
1245: and
1246: NASA grant HST-AR-09215 (RWR).
1247:
1248: This work is based on observations obtained at the
1249: Gemini Observatory, which is operated by the
1250: Association of Universities for Research in Astronomy, Inc.,
1251: under a cooperative agreement with the NSF on behalf of the
1252: Gemini partnership: the National Science Foundation (United States),
1253: the Particle Physics and Astronomy Research Council (United Kingdom),
1254: the National Research Council (Canada), CONICYT (Chile),
1255: the Australian Research Council (Australia), CNPq (Brazil)
1256: and CONICET (Argentina).
1257: Our observations were obtained with the Adaptive
1258: Optics System Hokupa'a/QUIRC, developed and operated by the
1259: University of Hawaii Adaptive Optics Group, with support from the
1260: National Science Foundation.
1261:
1262: %% Appendix material should be preceded with a single \appendix command.
1263: %% There should be a \section command for each appendix. Mark appendix
1264: %% subsections with the same markup you use in the main body of the paper.
1265: %% Each Appendix (indicated with \section) will be lettered A, B, C, etc.
1266: %% The equation counter will reset when it encounters the \appendix
1267: %% command and will number appendix equations (A1), (A2), etc.
1268:
1269: %\appendix
1270:
1271: %\section{Appendicial material}
1272:
1273: %% The reference list follows the main body and any appendices.
1274: %% Use LaTeX's thebibliography environment to mark up your reference list.
1275: %% Note \begin{thebibliography} is followed by an empty set of
1276: %% curly braces. If you forget this, LaTeX will generate the error
1277: %% "Perhaps a missing \item?".
1278: %%
1279: %% thebibliography produces citations in the text using \bibitem-\cite
1280: %% cross-referencing. Each reference is preceded by a
1281: %% \bibitem command that defines in curly braces the KEY that corresponds
1282: %% to the KEY in the \cite commands (see the first section above).
1283: %% Make sure that you provide a unique KEY for every \bibitem or else the
1284: %% paper will not LaTeX. The square brackets should contain
1285: %% the citation text that LaTeX will insert in
1286: %% place of the \cite commands.
1287: %%
1288: %% We have used macros to produce journal name abbreviations.
1289: %% AASTeX provides a number of these for the more frequently-cited journals.
1290: %% See the Author Guide for a list of them.
1291: %%
1292: %% Note that the style of the \bibitem labels (in []) is slightly
1293: %% different from previous examples. The natbib system solves a host
1294: %% of citation expression problems, but it is necessary to clearly
1295: %% delimit the year from the author name used in the citation.
1296: %% See the natbib documentation for more details and options.
1297:
1298: % REFERENCES
1299:
1300: \bibliographystyle{apj}
1301: \bibliography{psrwind}
1302:
1303: \clearpage
1304:
1305: %% Use the figure environment and \plotone or \plottwo to include
1306: %% figures and captions in your electronic submission.
1307:
1308: \begin{figure}
1309: \plotone{fig1.eps}
1310: \caption{
1311: Near-infrared image of the central
1312: $20\arcsec\times20\arcsec$ region of the Crab PWN,
1313: constructed from four dithered, $0.12\,{\rm ks}$ exposures
1314: in the $K'$ filter, taken on 2002 February 6 with a resolution
1315: of ${\rm FWHM}=0.\arcsec19$.
1316: The principal features of the pulsar wind termination shock
1317: and its environment are labeled,
1318: revealing the `crossbow' morphology characteristic of many PWN:
1319: arc-like, equatorial wisps and jet-like, polar knots
1320: symmetric about the projected rotation axis of the pulsar.
1321: Star I is the Hokupa'a guide star.
1322: Star II is another field star.
1323: \label{fig:gem1}}
1324: \end{figure}
1325:
1326: \begin{figure}
1327: \includegraphics[angle=90,height=6cm]{fig2a.ps}
1328: \includegraphics[angle=90,height=6cm]{fig2b.ps}
1329: %\includegraphics[angle=90,height=6cm]{histog1.ps}
1330: %\includegraphics[angle=90,height=6cm]{histogannulus.ps}
1331: \caption{
1332: Statistics of the nebula background.
1333: $a$. {\em Left.}
1334: Frequency histogram of pixel counts in a typical $K'$ image
1335: constructed from $4\times0.12{\rm ks}$ dithered exposures
1336: on 2002 February 6,
1337: {\em excluding} point sources and extended features.
1338: Raw counts (solid curve) and counts averaged over
1339: $50\times50$ pixel blocks (dashed curve) are shown;
1340: the curves overlap and are indistinguishable to the eye.
1341: $b$. {\em Right.}
1342: Frequency histogram of pixel counts in an annulus,
1343: radius $6.0{\rm FWHM}$ and one pixel thick,
1344: centered on the guide star (Star I).
1345: \label{fig:gem2}}
1346: \end{figure}
1347:
1348: \begin{figure}
1349: \plotone{fig3.eps}
1350: %\plotone{nt2jWcom_curves.eps}
1351: \caption{
1352: Enlarged $J$-band snapshots of the bright and faint wisps,
1353: taken on 2002 February 7, separated by $1.2\,{\rm ks}$.
1354: Orientation is as for Figure \ref{fig:gem1}, with north up
1355: and east left. Each image has been background-subtracted and
1356: normalized to the flux of Star I. The gray scale is the same
1357: in both images and is indicated by the central bar.
1358: The units on the bar indicate the normalized flux
1359: (defined as in Figure \ref{fig:gem4}) in a $1\arcsec\times 1\arcsec$ box.
1360: The apertures used to measure the total flux of the wisps,
1361: plotted in Figure \ref{fig:gem4}, are drawn as dotted curves.
1362: The seeing and the guide star flux are the same in the two exposures
1363: to within $0.\arcsec04$ and $4$ per cent respectively.
1364: \label{fig:gem3}}
1365: \end{figure}
1366:
1367: \begin{figure}
1368: \plotone{fig4.ps}
1369: %\plotone{wispsvst.ps}
1370: \caption{
1371: Light curves of the bright and faint wisps,
1372: after nebula subtraction and guide star normalization,
1373: as functions of time (in seconds)
1374: during the three nights of observations
1375: (demarcated by vertical lines).
1376: The four curves represent the
1377: bright wisp in $K'$ (solid),
1378: faint wisp in $K'$ (dashed),
1379: bright wisp in $J$ (dotted),
1380: and
1381: faint wisp in $J$ (dash-dotted).
1382: \label{fig:gem4}}
1383: \end{figure}
1384:
1385: \begin{figure}
1386: \plotone{fig5.eps}
1387: %plotone{nt2jScom_curves.eps}
1388: \caption{
1389: Enlarged $J$-band snapshots of the sprite (the fainter, larger
1390: feature on the left) and the rod (the brighter, smaller feature
1391: at center right),
1392: taken on 2002 February 7, separated by $1.2\,{\rm ks}$.
1393: Orientation is as for Figure \ref{fig:gem1}, with north up
1394: and east left. Each image has been
1395: background-subtracted and normalized to the flux of Star I. The gray
1396: scale is the same in both images and is indicated by the central bar.
1397: The units on the bar indicate the normalized flux (defined as in
1398: Figure \ref{fig:gem4}) in a $1\arcsec\times 1\arcsec$ box.
1399: The apertures used to measure the total
1400: fluxes of the sprite plus rod (plotted in Figure \ref{fig:gem6})
1401: and the rod alone (quoted in Table \ref{tab:gem1})
1402: are drawn as dotted polygons.
1403: The seeing and the guide star flux are the same in the two exposures
1404: to within $0.\arcsec04$ and $4$ per cent respectively.
1405: \label{fig:gem5}}
1406: \end{figure}
1407:
1408: \begin{figure}
1409: \plotone{fig6.ps}
1410: %\plotone{spritevst.ps}
1411: \caption{
1412: Light curve of the sprite,
1413: after nebula subtraction and guide star normalization,
1414: as a function of time (in seconds)
1415: during the three nights of observations
1416: (demarcated by vertical lines).
1417: The two curves represent $K'$ (solid) and $J$ (dotted)
1418: data respectively.
1419: \label{fig:gem6}}
1420: \end{figure}
1421:
1422: \begin{figure}
1423: \plotone{fig7.eps}
1424: %\plottwo{fig7a.eps}{fig7b.ps}
1425: %\plottwo{imgknot.eps}{surfknot.ps}
1426: \caption{
1427: Enlarged $K'$ image of the inner knot after subtraction of
1428: the pulsar PSF,
1429: plotted as a gray scale image with isophotes
1430: (solid contours) overlaid.
1431: The subtraction is imperfect because the PSF
1432: template is taken from the guide star, yet the PSF varies
1433: across the field of view (\S\ref{sec:gem2}).
1434: Three apertures used in our photometry of
1435: the inner knot are also shown:
1436: two circles of radius $0.75$ ${\rm FWHM}$
1437: and $5.0$ ${\rm FWHM}$, centered on the pulsar,
1438: and
1439: a circle of radius 10 pixels, centered on the brightness
1440: centroid of the inner knot.
1441: An annular region of inner radius $6.0$ ${\rm FWHM}$
1442: and one pixel width, also centered on the inner knot
1443: and used to estimate the background
1444: (\S\ref{sec:gem3c}),
1445: lies outside the borders of the image.
1446: \label{fig:gem7}}
1447: \end{figure}
1448:
1449: \begin{figure}
1450: \plotone{fig8.ps}
1451: \caption{
1452: Light curve of the inner knot,
1453: plotted as the flux ratio difference
1454: $I_{\rm k}/I_{\rm g}=
1455: I_{\rm p}(r_2)/I_{\rm g}(r_2) -
1456: I_{\rm p}(r_1)/I_{\rm g}(r_1)$
1457: for apertures with $r_1=0.75$ ${\rm FWHM}$ and $r_2=5.0$ ${\rm FWHM}$.
1458: The two curves represent $K'$ (solid) and $J$ (dotted)
1459: data respectively.
1460: \label{fig:gem8}}
1461: \end{figure}
1462:
1463: \begin{figure}
1464: \plotone{fig9.ps}
1465: \caption{
1466: Light curve of the inner knot,
1467: plotted as the flux enclosed by a circular aperture
1468: of radius 10 pixels
1469: (after nebula subtraction and guide star normalization).
1470: This aperture includes some leakage from the pulsar PSF.
1471: The two curves represent $K'$ (solid) and $J$ (dotted)
1472: data respectively.
1473: \label{fig:gem9}}
1474: \end{figure}
1475:
1476: \clearpage
1477:
1478: \begin{table}
1479: \begin{tabular}{lcccc} \hline
1480: Point source &$J$ flux &$K'$ flux &$\alpha$ &length\\
1481: &[mJy] &[mJy] & &[arcsec]\\
1482: \hline
1483: Star I &6.25 (0.12) &3.09 (0.07) &$ 1.24$ &---\\
1484: Star II &1.01 (0.02) &0.76 (0.02) &$ 0.50$ &---\\
1485: Pulsar &3.21 (0.15) &2.62 (0.17) &$ 0.36$ &---\\
1486: Knot &0.32 (0.16) &0.49 (0.17) &$-0.78$ &---\\
1487: \hline
1488: Extended feature &$J$ flux &$K'$ flux &$\alpha$ &length\\
1489: &[$\mu$Jy arcsec$^{-1}$] &[$\mu$Jy arcsec$^{-1}$] & &[arcsec]\\
1490: \hline
1491: Bright wisp &69.4 (1.4) &97.7 (2.2) &$-0.60$ &7.3\\
1492: Faint wisp &59.8 (1.3) &80.3 (1.8) &$-0.52$ &3.3\\
1493: Rod &62.8 (1.3) &68.7 (1.6) &$-0.16$ &1.2\\
1494: Sprite &46.2 (1.0) &53.5 (1.2) &$-0.26$ &2.5\\
1495: \hline
1496: \end{tabular}
1497: \caption{De-reddened fluxes and power-law spectral indices
1498: ($F_\nu \propto \nu^\alpha$)
1499: for features present in the images,
1500: collected into point sources (including the inner knot; top)
1501: and extended features (bottom).
1502: Fluxes for the extended features are quoted per arcsecond of length;
1503: the lengths are shown in the right-hand column.
1504: Errors ($1\sigma$) are quoted in parentheses after the fluxes.
1505: A correction has been made for interstellar extinction (Eikenberry et al.\ 1997).
1506: \label{tab:gem1}}
1507: \end{table}
1508:
1509: %% If you are not including electonic art with your submission, you may
1510: %% mark up your captions using the \figcaption command. See the
1511: %% User Guide for details.
1512: %%
1513: %% No more than seven \figcaption commands are allowed per page,
1514: %% so if you have more than seven captions, insert a \clearpage
1515: %% after every seventh one.
1516:
1517: %% Tables should be submitted one per page, so put a \clearpage before
1518: %% each one.
1519: %% Two options are available to the author for producing tables: the
1520: %% deluxetable environment provided by the AASTeX package or the LaTeX
1521: %% table environment. Use of deluxetable is preferred.
1522: %% If you use the table environment, please indicate horizontal rules using
1523: %% \tableline, not \hline.
1524: %% Do not put multiple tabular environments within a single table.
1525: %% The optional \label should appear inside the \caption command.
1526:
1527: %% The following command ends your manuscript.
1528: %% LaTeX will ignore any text that appears after it.
1529:
1530: \end{document}
1531: