1: \documentclass{piner}
2: %\documentclass[12pt,preprint]{aastex}
3: \usepackage{rotating}
4: \begin{document}
5:
6: \shorttitle{Parsec-Scale Jets of TeV Blazars}
7:
8: \title{The Parsec-Scale Jets of the TeV Blazars H\,1426+428, 1ES\,1959+650, and PKS\,2155$-$304: 2001--2004}
9:
10: \author{B.~Glenn~Piner\altaffilmark{1}, Niraj~Pant\altaffilmark{1}, and Philip~G.~Edwards\altaffilmark{2}}
11:
12: \altaffiltext{1}{Department of Physics and Astronomy, Whittier College,
13: 13406 E. Philadelphia Street, Whittier, CA 90608; gpiner@whittier.edu}
14:
15: \altaffiltext{2}{CSIRO Australia Telescope National Facility, Locked Bag 194, Narrabri NSW 2390, Australia; Philip.Edwards@csiro.au}
16:
17: \begin{abstract}
18: We present Very Long Baseline Array (VLBA) observations of the TeV
19: blazars H\,1426+428, 1ES\,1959+650, and PKS\,2155$-$304 obtained during
20: the years 2001 through 2004. We observed H\,1426+428 at four epochs
21: at 8\,GHz, and found that its parsec-scale structure consisted of a
22: $\sim 17$\,mJy core and a single $\sim 3$\,mJy jet component with an
23: apparent speed of $2.09c\pm0.53c$. The blazar 1ES\,1959+650 was
24: observed at three epochs at frequencies of 15 and 22\,GHz. Spectral
25: index information from these dual-frequency observations was used to
26: definitively identify the core of the parsec-scale structure.
27: PKS\,2155$-$304 was observed at a single epoch at 15\,GHz with
28: dual-circular polarization, and we present the first VLBI
29: polarimetry image of this source.
30: %The fractional polarization at the
31: %position of the core was 3\%, and the electric vector position angle
32: %was $131\arcdeg$, about $30\arcdeg$ misaligned from the innermost
33: %jet position angle.
34: For 1ES\,1959+650 and PKS\,2155$-$304, the
35: current observations are combined with the VLBA observations from
36: our earlier paper to yield improved apparent speed measurements for
37: these sources with greatly reduced measurement errors. The new
38: apparent speed measured for component C2 in 1ES\,1959+650 is
39: $0.00c\pm0.04c$ (stationary), and the new apparent speed measured
40: for component C1 in PKS\,2155$-$304 is $0.93c\pm0.31c$.
41: We combine the new apparent speed measurements from this paper
42: with the apparent speeds measured in TeV blazar jets from our earlier papers to form
43: a current set of apparent speed measurements in TeV HBLs.
44: The mean peak apparent pattern speed in the jets of the TeV HBLs is about 1$c$. We conclude
45: the paper with a detailed discussion of the interpretation of the
46: collected VLBA data on TeV blazars in the context of current
47: theoretical models for the parsec-scale structure of TeV blazar
48: jets.
49: \end{abstract}
50:
51: \keywords{
52: BL Lacertae objects: individual (H\,1426+428, 1ES\,1959+650, PKS\,2155$-$304) ---
53: galaxies: active ---
54: galaxies: jets --- radio continuum: galaxies}
55:
56: \section{Introduction}
57: \label{intro}
58:
59: The field of TeV gamma-ray astronomy has grown rapidly over the past
60: several years with the advent of sensitive new gamma-ray
61: telescopes such as H.E.S.S.\ and MAGIC. Particularly interesting
62: results have been obtained in blazar astronomy, as the total number of
63: detected TeV blazars has increased from six only a few years ago to 17
64: this year (Wagner 2007b). The majority of the detected TeV blazars
65: (16 of 17) belong to the class of high-frequency peaked BL Lac
66: objects, or HBLs --- so named because the two peaks in their spectral
67: energy distribution (SED) occur at relatively high UV/X-ray and
68: GeV/TeV energies. This two-peaked spectral energy distribution in HBLs
69: is most commonly interpreted as the result of relativistic electrons
70: (and possibly positrons) radiating in a jet which is undergoing bulk
71: relativistic motion at a small angle to the observer's line of sight,
72: with the low-frequency peak due to synchrotron radiation, and the
73: high-frequency peak due to inverse-Compton scattering of the jet's own
74: synchrotron photons (synchrotron self-Compton, or SSC emission). The
75: TeV source list is mostly limited to relatively nearby blazars ($z\lesssim0.2$),
76: because of the absorption of TeV gamma-rays on the extragalactic
77: background light.
78:
79: The TeV observations have shown dramatic variability in a number of
80: these blazars. Among the most remarkable variability events are the
81: 200\,s variability timescale detected for PKS 2155$-$304 in July 2006
82: by H.E.S.S.\ (Aharonian et al.\ 2007), and the 3-minute variations
83: detected for Mkn\,501 in June 2005 by MAGIC (Wagner 2007a). Such
84: rapid variations suggest extremely small emitting volumes and/or time
85: compression by large relativistic Doppler factors, e.g., $\delta\gtrsim100$
86: for PKS 2155$-$304 (Aharonian et al.\ 2007).
87: High Doppler factors are also sometimes invoked in specific SSC models
88: of TeV blazar spectra and variability; for example, Fossati et al. (2007)
89: consider two SSC models to explain the X-ray/TeV variability of Mkn 421:
90: one with $\delta\sim20$ (scattering in the Klein-Nishina regime), and
91: one with $\delta\sim100$ (scattering in the Thomson regime).
92: Such high Doppler factors are
93: at the upper limit of what is expected in relativistic jets, and
94: challenge our understanding of these objects if they do in fact occur.
95:
96: Complementary observations that are crucial to unraveling the physics
97: of TeV blazar jets are provided by the VLBI technique, which yields
98: radio images of the relativistic jets with sub-parsec resolutions for
99: these nearby blazars. Apparent jet speeds, brightness temperatures,
100: and limits on jet/counterjet brightness ratios can all be measured
101: from VLBI images, and these quantities all provide constraints on
102: fundamental jet parameters such as the bulk Lorentz factor and the
103: angle of the jet to the line-of-sight (subject to some caveats
104: discussed at length in $\S$~\ref{disc}). Our understanding of the
105: parsec-scale radio properties of the HBL class in general has been
106: increased by the recent work of Giroletti et al.\ (2004a, 2006). Those
107: authors studied a sample of low-redshift BL Lac objects with a variety
108: of radio instruments, and demonstrated that the radio properties of
109: the HBLs are consistent with them being the beamed versions of nearby
110: low-luminosity FR I radio galaxies. The TeV blazars thus likely have
111: an intrinsically different parent population with weaker jets,
112: compared to the more distant powerful blazars, which likely have FR II
113: parents (e.g., Urry \& Padovani 1995). The mean derived parsec-scale
114: Lorentz factor for the HBL class by Giroletti et al.\ (2004a),
115: including TeV sources, is only $<\Gamma>\sim3$, much lower than the
116: Lorentz factors suggested by the TeV gamma-ray emission.
117:
118: We have previously published a series of papers investigating the
119: parsec-scale kinematic properties of TeV-detected HBLs through
120: multi-epoch high-resolution VLBI observations, predominantly
121: with the National Radio
122: Astronomy Observatory's Very Long Baseline Array (VLBA) \footnote{The
123: National Radio Astronomy Observatory is a facility of the National
124: Science Foundation operated under cooperative agreement by
125: Associated Universities, Inc.}, and we extend that study in this
126: paper. Previous papers in this series have included: early VLBA and
127: space VLBI
128: observations of Mkn\,421 (Piner et al.\ 1999),
129: VLBI observations of Mkn\,501 (Edwards \& Piner 2002),
130: VLBA observations of 1ES\,1959+650,
131: PKS\,2155$-$304, and 1ES\,2344+514 (Piner \& Edwards 2004), and new VLBA
132: polarimetry observations of Mkn\,421 (Piner \& Edwards 2005). This
133: fifth paper in the series adds new VLBA observations of the TeV
134: blazars \object{1ES 1959+650} and \object{PKS 2155-304} whose
135: kinematics were relatively poorly determined in Piner \& Edwards
136: (2004) (hereafter Paper~I), as well as a four-epoch series of VLBA
137: observations of \object{H 1426+428} --- the sixth TeV blazar to be
138: discovered. Altogether we present six new images of 1ES\,1959+650,
139: four new images of H\,1426+428, and one new polarization image of
140: PKS\,2155$-$304, for a total of eleven new datasets. These datasets
141: comprise the observations from our TeV blazar monitoring program over
142: the years 2001-2004 (excepting the observations of Mkn\,421, which were
143: published separately in Piner \& Edwards [2005]). Note that except
144: for the two brightest sources (Mkn\,421 and Mkn\,501), the TeV blazars
145: are too faint in the radio ($\lesssim$100\,mJy) to be included in other
146: VLBA monitoring programs such as MOJAVE, and they typically require
147: long observations to obtain images of sufficient dynamic range.
148: Observational backgrounds on the three specific sources studied in
149: this paper are presented in the results section for the specific
150: source.
151:
152: Our earlier work on the parsec-scale structure of TeV blazars has
153: shown a noticeable lack of superluminal components in their jets,
154: which contrasts with the rapid superluminal motions observed in the
155: jets of more powerful blazars, and with the high Lorentz factors
156: derived from modeling the TeV emission.
157: %(An exception to this trend
158: %is BL~Lac, recently detected as a TeV source
159: %[Hayashida et al.\ 2007]
160: %and with superluminal speeds having been measured for multiple jet components
161: %[Mutel et al.\ 1990; Kellermann et al.\ 2004], however BL~Lac is a
162: %low-frequency peaked BL Lac object (LBL) and
163: %differing properties to HBLs are not unexpected.
164: We have previously
165: interpreted the general lack of superluminal components in HBLs
166: as evidence for a lower bulk Lorentz factor in the
167: parsec-scale radio-emitting region compared to the TeV-emitting
168: region. Models that have been invoked to explain this `bulk Lorentz
169: factor crisis' include: jets that are decelerated along their length
170: (Georganopoulos \& Kazanas 2003; Wang, Li, \& Xue 2004; Bicknell et
171: al.\ 2005), jets with transverse velocity structures consisting of a
172: fast spine and a slower layer (Giroletti et al.\ 2004b; Ghisellini,
173: Tavecchio, \& Chiaberge 2005; Henri \& Saug\'{e} 2006), or jets with
174: opening angles large enough that unintentional averaging over multiple
175: viewing angles (because of limited resolution) causes the apparent
176: conflict (Gopal-Krishna, Dhurde, \& Wiita 2004; Gopal-Krishna, Wiita,
177: \& Dhurde 2006). Finally, Gopal-Krishna et al.\ (2007) consider a
178: combination of the last two models; i.e., large opening angle jets with
179: transverse velocity structures. Some of these models may be
180: distinguished through the observed statistical distribution of
181: apparent speeds in TeV blazar jets (e.g., Gopal-Krishna et al.\ 2006),
182: so we conclude the paper by presenting our current best set of TeV
183: blazar apparent speed measurements from the complete series of five
184: papers, which consists of sixteen component speeds in six sources, and
185: discussing the various models in the context of the current
186: observations.
187:
188: In this paper we use cosmological parameters $H_{0}=71$ km s$^{-1}$
189: Mpc$^{-1}$, $\Omega_{m}=0.27$, and $\Omega_{\Lambda}=0.73$ (Bennett et
190: al.\ 2003). When results from other papers are quoted, they have been
191: converted to this cosmology.
192:
193: \section{Observations and Data Reduction}
194: \label{obs}
195:
196: We observed H\,1426+428 with the VLBA at 8.4\,GHz at four epochs
197: between 2001 July and 2003 October, under observation codes BE024 and
198: BE029.
199: A VLBI observation of H\,1426+428 had previously been made at 5\,GHz
200: by Kollgaard, Gabuzda, \& Feigelson (1996), however our
201: first observation (BE024 on 2001 July 4) was a pilot study
202: to see if H\,1426+428 had a parsec-scale jet that could be imaged
203: at 8\,GHz with
204: the VLBA, and to determine a more accurate position for
205: the source. That observation was carried out in
206: phase-referencing mode both to determine a milliarcsecond (mas) level
207: position, and because the correlated flux density of
208: H\,1426+428 was uncertain. The
209: source J1419+3821, for which a sub-mas position has been
210: determined by the International Celestial Reference Frame
211: (Ma et al.\ 1998), was used as the phase-reference source.
212: J1419+3821 has an angular separation of 4.6$^\circ$ from
213: H\,1426+428.
214:
215: \begin{table*}[!t]
216: \caption{Observation Log}
217: \begin{center}
218: \label{obstab}
219: {\scriptsize \begin{tabular}{l l l l c c c} \tableline \tableline \\ [-5pt]
220: & & & & & Time On & \\
221: & & \multicolumn{1}{c}{VLBA Observing} & \multicolumn{1}{c}{VLBA} & Frequency & Source & \\
222: \multicolumn{1}{c}{Source} & \multicolumn{1}{c}{Epoch} & \multicolumn{1}{c}{Code} &
223: \multicolumn{1}{c}{Antennas$^{a}$} & (GHz) & (hr) & Dual Pol$^{b}$ \\ \tableline \\ [-5pt]
224: H\,1426+428 & 2001 Jul 4 & BE024 & All & 8.4 & 7 & No \\
225: & 2003 Mar 22 & BE029A & No SC & 8.4 & 7 & No \\
226: & 2003 Jul 6 & BE029B & All & 8.4 & 7 & No \\
227: & 2003 Oct 19 & BE029C & All & 8.4 & 7 & No \\
228: 1ES\,1959+650 & 2003 Jun 9 & BE030A & No KP & 15.4 & 2.5 & No \\
229: & 2003 Jun 9 & BE030A & No KP & 22.2 & 2.5 & No \\
230: & 2003 Oct 31 & BE030B & All & 15.4 & 2.5 & No \\
231: & 2003 Oct 31 & BE030B & All & 22.2 & 2.5 & No \\
232: & 2004 Feb 8 & BE030C & All & 15.4 & 2.5 & No \\
233: & 2004 Feb 8 & BE030C & All & 22.2 & 2.5 & No \\
234: PKS\,2155$-$304 & 2003 Sep 20 & BE031 & No BR & 15.4 & 4 & Yes \\ \tableline
235: \end{tabular}}
236: \end{center}
237: {\scriptsize $a$: BR = Brewster, Washington,
238: KP = Kitt Peak,Arizona, and SC = Saint Croix, U.S. Virgin Islands.}\\
239: {\scriptsize $b$: Whether or not the experiment recorded dual-circular polarization.}
240: \end{table*}
241:
242: \begin{table*}
243: \caption{Parameters of the Images}
244: \begin{center}
245: \label{imtab}
246: {\scriptsize \begin{tabular}{l l c l c c l c c} \tableline \tableline \\ [-5pt]
247: & & & \multicolumn{3}{c}{\underline{Natural Weighting}} & \multicolumn{3}{c}{\underline{Uniform Weighting}} \\ [5pt]
248: & & & & Peak Flux & Lowest & & Peak Flux & Lowest \\
249: & & Frequency & & Density & Contour$^{b}$ & & Density & Contour$^{b}$ \\
250: \multicolumn{1}{c}{Source} & \multicolumn{1}{c}{Epoch} & (GHz) & \multicolumn{1}{c}{Beam$^{a}$} &
251: (mJy beam$^{-1}$) & (mJy beam$^{-1}$) & \multicolumn{1}{c}{Beam$^{a}$}
252: & (mJy beam$^{-1}$) & (mJy beam$^{-1}$) \\ \tableline \\ [-5pt]
253: H\,1426+428 & 2001 Jul 4 & 8.4 & 1.54,1.10,$-$4.9 & 20 & 0.13 & 1.07,0.77,$-$7.4 & 18 & 0.28 \\
254: & 2003 Mar 22 & 8.4 & 2.17,1.12,$-$19.2 & 19 & 0.12 & 1.63,0.75,$-$20.7 & 19 & 0.22 \\
255: & 2003 Jul 6 & 8.4 & 1.80,1.01,$-$7.9 & 16 & 0.11 & 1.30,0.69,$-$8.7 & 15 & 0.20 \\
256: & 2003 Oct 19 & 8.4 & 1.77,1.01,$-$5.2 & 16 & 0.11 & 1.26,0.69,$-$4.4 & 15 & 0.19 \\
257: 1ES\,1959+650 & 2003 Jun 9 & 15.4 & 0.97,0.48,$-$4.3 & 107 & 0.37 & 0.74,0.36,$-$2.4 & 98 & 0.67 \\
258: & 2003 Oct 31 & 15.4 & 1.01,0.53,$-$6.1 & 104 & 0.35 & 0.75,0.38,$-$5.2 & 95 & 0.55 \\
259: & 2004 Feb 8 & 15.4 & 1.12,0.54,5.3 & 77 & 0.36 & 0.77,0.36,1.3 & 67 & 0.70 \\
260: & 2003 Jun 9 & 22.2 & 0.65,0.32,$-$1.6 & 101 & 1.09 & 0.52,0.26,0.8 & 92 & 1.68 \\
261: & 2003 Oct 31 & 22.2 & 0.72,0.41,$-$1.0 & 84 & 0.78 & 0.53,0.26,$-$0.6 & 76 & 1.38 \\
262: & 2004 Feb 8 & 22.2 & 0.85,0.43,10.5 & 61 & 0.72 & 0.56,0.26,10.0 & 52 & 1.41 \\
263: PKS\,2155$-$304 & 2003 Sep 20 & 15.4 & 1.92,0.51,$-$11.0 & 122 & 0.38 & 1.30,0.36,$-$7.8 & 110 & 0.77 \\ \tableline
264: \end{tabular}}
265: \end{center}
266: %\\ [5pt]
267: {\scriptsize $a$: Numbers given for the beam are the FWHMs of the major
268: and minor axes in mas, and the position angle of the major axis in degrees.
269: Position angle is measured from north through east.}\\
270: {\scriptsize $b$: The lowest contour is set to be three times the rms noise
271: in the image. Successive contours are each a factor of 2 higher.}
272: \end{table*}
273:
274: We were able to model the phase-reference source, J1419+3821,
275: with two circular Gaussian components: a 0.34\,Jy unresolved core
276: component and a 0.16\,Jy
277: component with a FWHM size of 1.2\,mas located 0.6\,mas to the south.
278: Based on the ICRF position for J1419+3821, we
279: derived a (J2000) position for H\,1426+428 of
280: ($\alpha$,$\delta$)=(14h28m32.609s, +42$^\circ$40$'$21.05$''$),
281: which we expect to have an accuracy of better than 20\,mas.
282: With the use of this position,
283: phase-referencing was unnecessary in subsequent epochs,
284: as fringes could be directly detected to H\,1426+428,
285: with a correlated flux density of $\sim20$\,mJy at 8\,GHz.
286: Following the successful imaging of parsec-scale
287: structure in H\,1426+428 in the pilot study,
288: three further epochs were observed throughout
289: 2003 (observation code BE029) to study the jet kinematics. All four
290: observations recorded approximately 7 hours on-source, for an expected
291: thermal noise limit of about 0.05\,mJy beam$^{-1}$. Note that
292: H\,1426+428 was observed at a lower frequency than the other sources in
293: this paper and in Paper~I, in order to achieve the needed sensitivity
294: to image this fainter source.
295:
296: 1ES\,1959+650 was observed with the VLBA during three 6-hour sessions
297: between 2003 June and 2004 February, under observation code BE030.
298: The observations were split between 15.4 and 22.2\,GHz, with
299: approximately 2.5 hours spent on source at each frequency at each
300: epoch (with approximately one hour of calibration scans and slewing).
301: This source was observed at two frequencies in order to obtain
302: spectral index information between the 15 and 22\,GHz images, to help
303: resolve potential ambiguities in the source structure discussed in Paper~I. The
304: 15\,GHz data has been combined with the three epochs at this frequency
305: from Paper~I, to obtain kinematic information over a six-epoch series
306: for this source.
307:
308: PKS\,2155$-$304 was observed at a single 6-hour epoch on 2003 September
309: 20 at 15.4\,GHz, with dual-circular polarization, under observation
310: code BE031. The goal of this observation was two-fold: to provide an
311: additional epoch several years removed from the three epochs in Paper~I
312: to better constrain the jet proper motion (which was poorly
313: constrained in Paper~I), and to produce the first VLBI polarimetry
314: image of this source. Previous observations in Paper~I had shown the
315: correlated flux density was large enough (at $>$100\,mJy) to produce a
316: useful polarization image at this frequency. Approximately four hours
317: were spent on-source in this observation, with the additional two
318: hours spent on polarization calibration sources.
319:
320: \begin{table*}[!t]
321: \caption{Circular Gaussian Models}
322: \begin{center}
323: \label{mfittab}
324: {\scriptsize \begin{tabular}{l l c c c c r c c c} \tableline \tableline \\ [-5pt]
325: & & Frequency & & $S$ & $r$ &
326: \multicolumn{1}{c}{PA} & $a$ & & $T_{B}$ \\
327: \multicolumn{1}{c}{Source} & \multicolumn{1}{c}{Epoch} & (GHz) & Component & (mJy) & (mas) &
328: \multicolumn{1}{c}{(deg)} & (mas) & $\chi_{R}^{2}$ & ($10^{10}$ K) \\
329: \multicolumn{1}{c}{(1)} & \multicolumn{1}{c}{(2)} & (3) & (4) & (5) & (6) &
330: \multicolumn{1}{c}{(7)} & (8) & (9) & (10) \\ \tableline \\ [-5pt]
331: H\,1426+428 & 2001 Jul 4 & 8.4 & Core & 19 & ... & ... & 0.22 & 0.96 & 0.8 \\
332: & & & C1 & 4 & 1.02 & $-$21.4 & 1.38 & & \\
333: & 2003 Mar 22 & 8.4 & Core & 19 & ... & ... & 0.17 & 0.67 & 1.2 \\
334: & & & C1 & 3 & 1.65 & $-$26.8 & 1.46 & & \\
335: & 2003 Jul 6 & 8.4 & Core & 16 & ... & ... & 0.19 & 0.64 & 0.8 \\
336: & & & C1 & 3 & 1.60 & $-$25.5 & 1.74 & & \\
337: & 2003 Oct 19 & 8.4 & Core & 15 & ... & ... & 0.14 & 0.64 & 1.5 \\
338: & & & C1 & 4 & 1.48 & $-$23.9 & 2.03 & & \\
339: 1ES\,1959+650 & 2003 Jun 9 & 15.4 & Core & 81 & ... & ... & 0.07 & 0.72 & 10.0 \\
340: & & & C2 & 72 & 0.34 & 124.0 & 0.55 & & \\
341: & 2003 Oct 31 & 15.4 & Core & 81 & ... & ... & 0.09 & 0.72 & 5.9 \\
342: & & & C2 & 58 & 0.32 & 125.5 & 0.62 & & \\
343: & 2004 Feb 8 & 15.4 & Core & 63 & ... & ... & 0.15 & 0.73 & 1.5 \\
344: & & & C2 & 46 & 0.41 & 129.8 & 0.51 & & \\
345: & 2003 Jun 9 & 22.2 & Core & 90 & ... & ... & 0.09 & 0.58 & 2.6 \\
346: & & & C2 & 72 & 0.33 & 124.2 & 0.48 & & \\
347: & 2003 Oct 31 & 22.2 & Core & 82 & ... & ... & 0.13 & 0.59 & 1.2 \\
348: & & & C2 & 35 & 0.42 & 126.5 & 0.54 & & \\
349: & 2004 Feb 8 & 22.2 & Core & 59 & ... & ... & 0.15 & 0.55 & 0.7 \\
350: & & & C2 & 30 & 0.46 & 126.7 & 0.47 & & \\
351: PKS\,2155$-$304 & 2003 Sep 20 & 15.4 & Core & 130 & ... & ... & 0.23 & 0.68 & 1.4 \\
352: & & & C1 & 15 & 1.04 & 163.9 & 0.48 & & \\ \tableline
353: \end{tabular}}
354: \end{center}
355: %\\ [5pt]
356: NOTES.--- Col. (5): Flux density in millijanskys.
357: Col. (6) and (7): $r$ and PA are the polar coordinates of the
358: center of the component relative to the presumed core.
359: Position angle is measured from north through east.
360: Col. (8): $a$ is the Full Width at Half Maximum (FWHM) of the circular Gaussian
361: component.
362: Col. (9): The reduced chi-squared of the model fit.
363: Col. (10): The maximum source-frame brightness temperature of the circular Gaussian core component is given by
364: $T_{B}=1.22\times10^{12}\;\frac{S(1+z)}{a^{2}\nu^{2}}$~K,
365: where $S$ is the flux density of the Gaussian in Janskys,
366: $a$ is the FWHM of the Gaussian in mas,
367: $\nu$ is the observation frequency in\,GHz, and $z$ is the redshift.
368: \end{table*}
369:
370: An observation log of all new observations discussed in this paper is
371: given in Table~\ref{obstab}. All observations were recorded at a data
372: rate of 128 Mbps. The AIPS software package was used for calibration
373: and fringe-fitting of the correlated visibilities, and the
374: visibilities were edited and final CLEAN images were produced using
375: the Difmap software package. For the polarization experiment BE031,
376: calibration of the polarization response of the feeds was done with
377: LPCAL in AIPS. The required electric vector position angle (EVPA)
378: correction for BE031 was applied using CLCOR in AIPS, and was
379: determined from the observed EVPA of calibrator source J2136+006
380: (which had a relatively stable EVPA during 2003), compared with the
381: EVPA recorded for this source on the VLA/VLBA Polarization Calibration
382: Page\footnote{http://www.vla.nrao.edu/astro/calib/polar/},
383: interpolated to our frequency and epoch of observation. Parameters
384: for all images are given in Table~\ref{imtab}; individual images are
385: discussed in the results sections on the individual sources below.
386: All data sets are shown restored with both uniform weighting
387: (uvweight=2,0 in Difmap, improving resolution at the expense of
388: signal-to-noise ratio) and natural weighting (uvweight=0,$-$2 in
389: Difmap). None of the naturally weighted images have an rms noise
390: level that significantly exceeds the expected thermal noise limit.
391:
392: \begin{sidewaysfigure*}[t]
393: \includegraphics[scale=0.80,angle=-90.0]{f1.ps}
394: \caption{VLBA images of H\,1426+428 at 8.4\,GHz.
395: The top row shows the images obtained with natural weighting,
396: the bottom row the images obtained with uniform weighting.
397: The axes are labeled in milliarcseconds (mas).
398: Numerical parameters of the images are given in Table~\ref{imtab}.
399: The location of the center of the circular Gaussian (excluding the Gaussian representing the core)
400: that was fit to the visibilities is marked with an asterisk.
401: Parameters of the Gaussian models are given in Table~\ref{mfittab}.}
402: \end{sidewaysfigure*}
403:
404: After final calibration of the visibilities, circular Gaussian model
405: components were fit to the visibilities using the {\em modelfit} task
406: in Difmap. Such models allow the structure of the jet to be described
407: using only a few numerical parameters. The source structures in the
408: images in this paper are relatively simple (partly due to the high
409: resolution and limited dynamic range of these observations, and some
410: of these sources show more complex structures on lower-resolution
411: images that may not be well modeled as circular Gaussians), and in all
412: cases satisfactory fits to the visibilities were obtained using only
413: two circular Gaussian components (representing the core and one jet
414: component). The model fits are tabulated in Table~\ref{mfittab}, and
415: results for each source are discussed in the results sections on the
416: individual sources below. The reduced $\chi^{2}$ for all model fits
417: was under 1.0. Model fitting directly to the visibilities allows
418: sub-beam resolution to be obtained (e.g., Kovalev et al.\ 2005), and
419: components can be clearly identified in the model fitting even when
420: they appear blended with the core component in the CLEAN images.
421:
422: \begin{figure*}[t]
423: \begin{center}
424: \includegraphics[scale=0.35,angle=90.0]{f2.ps}
425: \caption{Distance from the core of the center of Gaussian component C1 in H\,1426+428 as
426: a function of time. The line is the least-squares fit to outward motion with constant speed.}
427: \end{center}
428: \end{figure*}
429:
430: As in Paper~I, we estimate that these full-track observations allow us
431: to fit positions of model component centers to within $\sim10$\% of a
432: uniform beam width, and this is the error assumed for subsequent
433: analysis (calculated by taking 10\% of the projection of the
434: elliptical beam FWHM onto a line joining the center of the core to the
435: center of the component).
436: Error bars much larger than this result in fits to
437: linear component motion that are so good as to be statistically
438: unlikely, confirming that this estimated error is reasonable.
439: Note that the low rms noise achieved by these long
440: integrations allows us to detect even the faintest component (the 3
441: mJy component in H\,1426+428) with a signal-to-noise ratio exceeding
442: 30:1.
443: Lower limits to the fitted core sizes (which yield upper limits to the
444: measured brightness temperatures), were determined using the DIFWRAP
445: program for model component error analysis (Lovell 2000).
446: These size lower limits range from zero (completely unresolved components)
447: to only about 20\% less than the best-fit size, and they are discussed in the
448: text on the core brightness temperature upper limits for each
449: individual source below.
450:
451: \section {Results on Individual Sources}
452: \label{results}
453:
454: \subsection{H\,1426+428}
455:
456: The detection in TeV gamma-rays of the HBL H\,1426+428 by the Whipple,
457: HEGRA, and CAT telescopes (Horan et al.\ 2002; Aharonian et al.\ 2002;
458: Djannati-Ata{\"i} et al.\ 2002) made it the sixth HBL source to be detected
459: in high-energy gamma-rays, and, at a redshift of $z=0.129$, it was the
460: most distant TeV HBL known at that time. Further analysis of the TeV
461: gamma-ray emission is given by Petry et al.\ (2002) and Aharonian et
462: al.\ (2003). Aharonian et al.\ (2003) and Costamante et al.\ (2003)
463: consider various models for the absorption of gamma-rays by the
464: extragalactic background light, and find that the data imply an
465: intrinsic TeV spectrum which dominates the broad-band spectral energy
466: distribution. X-ray observations indicate that the synchrotron peak
467: is sometimes located in excess of 100 keV (Costamante et al.\ 2001;
468: Wolter et al.\ 2007), and that the location of the peak is
469: significantly variable (Falcone, Cui, \& Finley 2004). SSC models
470: have been applied to the SED of H\,1426+428 by Wolter et al.\ (2007),
471: Costamante et al.\ (2003) (both using the one-zone homogeneous SSC
472: model of Ghisellini, Celotti, \& Costamante [2002]), and Kato,
473: Kusunose, \& Takahara (2006); all of these models give Doppler and
474: bulk Lorentz factors of about 20, and jet viewing angles of about
475: 2$\arcdeg$, but involve numerous assumptions.
476:
477: \begin{figure*}[p]
478: \begin{center}
479: \includegraphics[scale=0.50,angle=-90.0]{f3.ps}
480: \caption{VLBA images of 1ES\,1959+650 at 15.4\,GHz.
481: The top row shows the images obtained with natural weighting,
482: the bottom row the images obtained with uniform weighting.
483: The axes are labeled in milliarcseconds (mas).
484: Numerical parameters of the images are given in Table~\ref{imtab}.
485: The location of the center of the circular Gaussian (excluding the Gaussian representing the core)
486: that was fit to the visibilities is marked with an asterisk.
487: Parameters of the Gaussian models are given in Table~\ref{mfittab}.}
488: \end{center}
489: \end{figure*}
490:
491: \begin{figure*}
492: \begin{center}
493: \includegraphics[scale=0.50,angle=-90.0]{f4.ps}
494: \caption{VLBA images of 1ES\,1959+650 at 22.2\,GHz.
495: The top row shows the images obtained with natural weighting,
496: the bottom row the images obtained with uniform weighting.
497: The axes are labeled in milliarcseconds (mas).
498: Numerical parameters of the images are given in Table~\ref{imtab}.
499: The location of the center of the circular Gaussian (excluding the Gaussian representing the core)
500: that was fit to the visibilities is marked with an asterisk.
501: Parameters of the Gaussian models are given in Table~\ref{mfittab}.}
502: \end{center}
503: \end{figure*}
504:
505: VLA images of this source are shown by Laurent-Muehleisen et al.\
506: (1993) and Giroletti et al.\ (2004a). Laurent-Muehleisen et al.\ (1993)
507: present a 1.5\,GHz combined A- and C-configuration image that shows
508: poorly represented diffuse emission north of the core. Giroletti et
509: al.\ (2004a) detect the presence of a faint halo surrounding the
510: central core, with a northeast extension oriented at
511: P.A.$\sim50\arcdeg$, in their 1.4\,GHz A-array image.
512: H\,1426+428 has
513: been previously observed with VLBI by Kollgaard et al.\
514: (1996), and by Giroletti et al.\ (2006).
515: The 5\,GHz MkIII VLBI image
516: from 1991 by Kollgaard et al.\ (1996) shows a 2\,mJy component 1.2 mas
517: northeast (P.A.$\approx20\arcdeg$) of a 19\,mJy compact core. The
518: source is completely unresolved in the lower resolution 1.6\,GHz
519: European VLBI Network (EVN) image in Giroletti et al.\ (2006). Based on
520: the available VLA and VLBI data, Giroletti et al.\ (2006) estimate a
521: viewing angle of $\theta\sim20\arcdeg$ and a Doppler factor of
522: $\delta\sim3$, but those estimates are based on assumptions that may
523: only be valid in a statistical sense for samples of many blazars (such
524: as that the Lorentz factor $\Gamma\sim1/\theta$).
525:
526: \begin{figure*}[t]
527: \begin{center}
528: \includegraphics[scale=0.35,angle=0.0]{f5.ps}
529: \end{center}
530: \caption{Tapered 15\,GHz VLBA image of 1ES\,1959+650 from 2004 Feb 8.
531: The restoring beam is 3.39 by 2.3 mas at a position angle of 37$\arcdeg$.
532: The lowest contour is drawn at a flux density of 0.42\,mJy beam$^{-1}$
533: (three times the rms noise level). Successive contours are each a factor of 2 higher.
534: The peak flux density is 107\,mJy beam$^{-1}$.}
535: \end{figure*}
536:
537: Our four 8.4\,GHz VLBA images of H\,1426+428 are shown in Figure~1.
538: The top row of Figure~1 shows the naturally weighted images, while the
539: bottom row shows the uniformly weighted images. Parameters of the
540: images are given in Table~\ref{imtab}. The linear resolution of these
541: images at the distance of H\,1426+428 is approximately 2.3 pc
542: mas$^{-1}$. The images all show a compact core with a flux density of
543: $\sim$17\,mJy, and a faint jet extending to the northwest, with a
544: position angle of about $-25\arcdeg$. The position angle of the
545: extended structure seen at all four epochs is similar, and differs
546: significantly from the position angle of the structure measured by
547: Kollgaard et al.\ (1996) from their single 5\,GHz MkIII VLBI image from
548: 1991 of $\approx20\arcdeg$. No counterjet is detected; although,
549: because of the low flux density of the jet, the limit that can be
550: placed on the jet to counterjet brightness ratio is a rather low 30:1.
551: Although variability and different resolutions may
552: cause significant uncertainty, we can estimate
553: a total VLBI spectral index between 1.6 GHz (Giroletti et al.\ 2006), 5 GHZ
554: (Kollgaard et al.\ 1996) and 8.4 GHz (this paper), and obtain
555: $\alpha=-0.26+\-0.10$ ($S\propto\nu^{\alpha}$), a flat but not inverted spectral index.
556:
557: The structure of H\,1426+428 at all four epochs is well modeled by two
558: circular Gaussian components, representing the core and a single jet
559: component, designated C1, with a flux density of about 3\,mJy.
560: Parameters of the Gaussian models are given in Table~\ref{mfittab}.
561: The measured brightness temperature of the core is about
562: $1\times10^{10}$~K (see Table~\ref{mfittab}), but this component is
563: unresolved (the visibilities are almost as well fit by a point-like
564: component), so that this brightness temperature should be considered a
565: lower-limit. The jet component appears to move out between 2001 and
566: 2003, although no motion can be detected in the three 2003 epochs
567: alone. A linear fit to the separation of the jet component from the
568: core vs. time is shown in Figure~2. The fit yields a measured apparent
569: speed for the jet component of $2.09c\pm0.53c$. Interpretation of
570: model fitting data over only a few widely spaced epochs can lead to
571: uncertainties in component identification (e.g., Piner et al.\ 2007),
572: but for this source the structure is quite simple, minimizing the
573: chances of confusing multiple components. If this measured apparent
574: speed is taken as the bulk apparent speed of the
575: jet\footnote {the apparent jet speed
576: $\beta_{app}=\beta\sin\theta/(1-\beta\cos\theta)$, and the Doppler
577: factor $\delta=1/\Gamma(1-\beta\cos\theta)$, where $\Gamma$ is the
578: bulk Lorentz factor, $\beta=v/c$, and $\theta$ is the angle to the
579: line of sight},
580: and is combined with the viewing angle derived from SSC modeling of
581: about 2$\arcdeg$ (see above), then the bulk Lorentz factor of the jet
582: at the location of C1 is $\Gamma=5.6$, and the Doppler factor is
583: $\delta=11$.
584:
585: The difference of $45\arcdeg$ in the position angle of the jet
586: structure from that measured twelve years previously by Kollgaard et
587: al.\ (1996) is interesting. Changes in position angles of component
588: ejections have been well observed in other sources such as BL Lac, and
589: have been interpreted as periodic precession of the jet nozzle caused
590: by binary supermassive black holes by some authors (e.g., Stirling et
591: al.\ 2003, see also Mutel \& Denn 2005).
592: To find out whether or not such variations are occurring
593: in H\,1426+428 will have to wait for the ejection of more components,
594: because the two position angles measured twelve years apart are not
595: enough to place any constraints on such a motion. We note that we do
596: not detect any significant changes in the position angle of component
597: C1 itself over the two years of monitoring covered in this paper
598: (the change in the position angle of C1 of up to $5.4\arcdeg$ in Table~\ref{mfittab}
599: corresponds to a transverse distance of only about 0.15 mas, about the size of
600: the 1$\sigma$ positional error bars for this source).
601:
602: \begin{figure*}[t]
603: \begin{center}
604: \includegraphics[scale=0.40]{f6.eps}
605: \end{center}
606: \caption{Spectral index map of 1ES\,1959+650 between 15 and 22\,GHz, on 2003 Oct 31.
607: Contours are drawn from spectral indices of $-$1.5 to 0.5 in steps of 0.25.
608: Spectral index is calculated as $S\propto\nu^{\alpha}$, where $S$ is the flux density,
609: $\nu$ is the observing frequency, and $\alpha$ is the spectral index. The spectral index has
610: been plotted at all points where the flux density is at least four times the flux density of the lowest
611: contour on the images at both frequencies, i.e, at least three contours up on both relevant images.}
612: \end{figure*}
613:
614: \subsection{1ES\,1959+650}
615:
616: Observational background on the $z=0.047$ HBL 1ES\,1959+650 for results
617: published prior to 2004 has been discussed in Paper~I. Some more
618: recent high-energy observational results include a spectral analysis
619: of the 2002 TeV flares, including the ``orphan'' TeV flare in 2002
620: June (a TeV flare that occurred without a corresponding X-ray flare)
621: by Daniel et al.\ (2005), and an analysis of low-level TeV emission
622: detected by the MAGIC telescope in 2004 by Albert et al.\ (2006).
623: Multi-wavelength campaigns on 1ES\,1959+650 from 2002 and 2003 are
624: described by Krawczynski et al.\ (2004) and Gutierrez et al.\ (2006),
625: respectively. Both of these papers model the multi-wavelength SED of
626: 1ES\,1959+650 with a simple one-zone SSC model with a Doppler beaming
627: factor of $\delta=20$. Krawczynski et al.\ (2004) also discuss the
628: difficulty of modeling the orphan flare mentioned above with
629: conventional one-zone SSC models, and they suggest several
630: alternatives, including multiple-component SSC models, external
631: Compton models, and high-energy proton models. A specific model for
632: producing orphan flares from relativistic protons is described by
633: B\"{o}ttcher (2005), but may require unreasonably high jet powers
634: (B\"{o}ttcher 2005, erratum). A multiple-component leptonic model
635: that succeeds in producing orphan flares is described by Kusunose and
636: Takahara (2006), using an inhomogeneous conical jet geometry, with the
637: line-of-sight inside the opening angle of the jet.
638:
639: On the parsec scale, the 5\,GHz VLBA images of 1ES\,1959+650 by Rector,
640: Gabuzda, \& Stocke (2003) and Bondi et al.\ (2001, 2004) all show a
641: diffuse jet with a broad ($\sim55\arcdeg$) opening angle extending 20
642: mas north of the core along a position angle of $\approx-5\arcdeg$.
643: VLA images by Rector et al.\ (2003) and Giroletti et al.\ (2004a) show
644: faint extended flux to both the north (P.A. $\approx-5\arcdeg$) and
645: south (P.A. $\approx175\arcdeg$) of the core, suggesting negligible
646: Doppler boosting by the time the jet has reached the arcsecond scale.
647: Giroletti et al.\ (2006) estimate a viewing angle of
648: $\theta\sim20\arcdeg$ and a Doppler factor of $\delta\sim3$ on the
649: parsec scale from the available radio data, similar to their result
650: for H\,1426+428 discussed above.
651:
652: Our three epoch series of 15\,GHz VLBA images of 1ES\,1959+650 from
653: 2003-2004 is shown in Figure~3, and the series of 22\,GHz images from
654: the same epochs is shown in Figure~4. Both sets of images are shown
655: restored with both uniform and natural weighting. Parameters of these
656: images are given in Table~\ref{imtab}. At the redshift of 1ES\,1959+650
657: ($z=0.047$), the linear scale of the images is 0.9 pc mas$^{-1}$.
658: These images can then be compared with our previous sequence of three
659: 15\,GHz VLBA images from 2000, as shown in Paper~I. The 15 and
660: 22\,GHz images show a similar parsec-scale morphology, that is also
661: similar to that seen for this source in Paper~I, but which is markedly
662: different from that seen in the lower resolution 5\,GHz VLBA images by
663: Rector et al.\ (2003) and Bondi et al.\ (2001, 2004). Figures~3 and 4
664: show a brighter, compact northwestern feature, with a short extension
665: to the southeast along a position angle of $\approx125\arcdeg$ ---
666: this is particularly evident in the uniformly weighted 22\,GHz images.
667: This extension matches well the location of component C2 seen in Paper~I.
668: The diffuse northern jet seen in the 5\,GHz VLBA images is only
669: faintly present in the naturally-weighted 15\,GHz images. If we apply
670: a taper to our 15\,GHz VLBA data, then we better detect the diffuse
671: northern emission. Figure~5 shows a tapered image from our 15\,GHz
672: observation on 2004 Feb 8; parts of the northern jet are clearly
673: visible, although at relatively low significance.
674:
675: The morphology is intriguing because of the extreme misalignment
676: between the parsec-scale structures seen at 5\,GHz on the one hand,
677: and 15 and 22\,GHz on the other. If the northwestern component on our
678: images is the core (assumed in Paper~I), then the jet starts to the
679: southeast with a position angle of $\approx125\arcdeg$ at 0.5 mas from
680: the core, before bending to a position angle of $\approx-5\arcdeg$ at
681: about 5 mas from the core. Such an extreme parsec-scale misalignment
682: would most likely be due to a smaller intrinsic bend, amplified by
683: projection effects caused by a small angle to the line-of-sight.
684: Alternatively, the fainter southeastern feature could be the core,
685: with the northwestern feature representing a brighter jet component,
686: as has been seen in 4C~39.25 (e.g., Alberdi et al.\ 1993).
687: Definitively identifying the core requires spectral information, which
688: is obtained here from the dual-frequency 15 and 22\,GHz observations.
689: The core is expected to have an inverted or nearly flat spectral index
690: due to synchrotron self-absorption, while the jet components are
691: expected to have optically thin spectral indices.
692:
693: A parsec-scale spectral index map of 1ES\,1959+650 for the 2003 Oct 31
694: epoch is shown in Figure~6. The other two epochs yield similar
695: spectral index maps. The map was constructed by restoring the 22 and
696: 15\,GHz images from that date with the same beam (using the average of
697: the naturally-weighted beams at each frequency), and calculating the
698: spectral index at each pixel. The sign convention used is
699: $S\propto\nu^{\alpha}$, where $S$ is the flux density, $\nu$ is the
700: observing frequency, and $\alpha$ is the spectral index. The
701: northwest corner of the source has the flattest spectral index, with
702: the spectral index steepening to the southeast, with a possible small
703: rise again at the location of the stationary component C2. The
704: spectral index map thus identifies the core as the brighter component
705: at the northwest end of the source, as was assumed in Paper~I, a
706: result confirmed by the dual-frequency model fitting described below.
707: This leaves a jet with an $\approx130\arcdeg$ apparent bend from the
708: southeast to the north between 15\,GHz and 5\,GHz VLBI scales as the
709: most viable interpretation of the parsec-scale morphology.
710:
711: The source visibilities are well modeled by two circular Gaussians at
712: both frequencies; the parameters of these models are given in
713: Table~\ref{mfittab}. The measured brightness temperature of the core
714: component is a few times $10^{10}$~K, and it is partially resolved
715: only at the final epoch at both frequencies, with a brightness
716: temperature upper limit of $6\times10^{10}$~K at both frequencies at
717: that epoch. This is the same as the brightness temperature upper
718: limit obtained for this source in Paper~I. The jet component is
719: blended with the core component on the CLEAN images, but because of
720: its relatively high flux density relative to the core, it is highly
721: significant in the model fitting. The location of this component
722: matches well the expected location of the component labeled C2 in
723: Paper~I, so this component is subsequently referred to as C2 (but note
724: that the same caveats discussed in the section on H\,1426+428 about the
725: interpretation of model fitting data over widely spaced epochs still
726: apply). The component labeled C1 in Paper~I that appeared south of C2
727: at a separation of about 0.8 mas from the core and a position angle of
728: 160$\arcdeg$ is no longer detected, presumably it has faded between
729: 2001 and 2003. Interestingly, a larger and diffuse Gaussian component
730: almost on top of the core is suggested by the model fitting, but not
731: at a very significant level (so it is not included in Table~\ref{mfittab}),
732: but we speculate that this may represent the older component C1
733: crossing almost over the core as the jet bends to the north.
734:
735: \begin{figure*}[p]
736: \begin{center}
737: \includegraphics[scale=0.40,angle=90.0]{f7.ps}
738: \caption{Distance from the core of the center of Gaussian component C2 in 1ES\,1959+650 as
739: a function of time. The line is the least-squares fit to outward motion with constant speed.}
740: \end{center}
741: \end{figure*}
742:
743: \begin{figure*}
744: \begin{center}
745: \includegraphics[scale=0.425]{f8.ps}
746: \end{center}
747: \caption{VLBA images of PKS\,2155$-$304 at 15.4\,GHz.
748: The top image is shown using natural weighting,
749: the bottom is shown using uniform weighting.
750: Numerical parameters of the images are given in Table~\ref{imtab}.
751: The location of the center of the circular Gaussian (excluding the Gaussian representing the core)
752: that was fit to the visibilities is marked with a diamond.
753: The tick marks show the magnitude of the polarized flux (with a scale of 0.2 mas mJy$^{-1}$
754: in the top image, and 0.1 mas mJy$^{-1}$ in the bottom image) and the direction of the EVPA.
755: Tick marks are drawn at pixels where the polarized flux is greater
756: than three times the rms noise in the polarization image.
757: The colors show the fractional polarization, with the scale indicated to the right of the images.}
758: \end{figure*}
759:
760: \begin{figure*}[t]
761: \begin{center}
762: \includegraphics[scale=0.40,angle=0.0]{f9.ps}
763: \end{center}
764: \caption{Tapered 15\,GHz VLBA image of PKS\,2155$-$304 from 2003 Sep 20.
765: The restoring beam is 4.38 by 1.44 mas at a position angle of $-4\arcdeg$.
766: The lowest contour is drawn at a flux density of 0.45\,mJy beam$^{-1}$
767: (three times the rms noise level). Successive contours are each a factor of 2 higher.
768: The peak flux density is 140\,mJy beam$^{-1}$.}
769: \end{figure*}
770:
771: The spectral indices of the two components can also be calculated from
772: these dual-frequency model fits. We average the spectral indices
773: calculated at all three epochs, and obtain an average spectral index
774: for the core of $0.04\pm0.18$, and an average spectral index for C2 of
775: $-0.85\pm0.18$, confirming the results obtained from the spectral
776: index map. The errors on the spectral indices were calculated from an
777: estimated 7\% flux error at 15\,GHz and an estimated 9\% flux error at
778: 22\,GHz, where these estimated flux errors were obtained from the
779: average correction applied by the {\em gscale} amplitude
780: self-calibration routine in Difmap to the antenna amplitude scales.
781: The average separation of component C2 from the core is slightly
782: larger at 22\,GHz than its average separation at 15\,GHz (by
783: $0.05\pm0.04$ mas), a result expected from inhomogeneous models of the
784: radio core, where the `core' lies farther back along the jet at higher
785: frequencies (e.g., K\"{o}nigl 1981). This measured
786: frequency-dependent separation was compensated for in the spectral
787: index map shown in Figure~6 by applying a relative shift in the proper
788: direction to align component C2 between the two frequencies.
789:
790: A linear fit to the separation of the jet component C2 from the core
791: vs. time from the six total 15.4\,GHz images (three from this paper,
792: three from Paper~I) is shown in Figure~7. The fit yields a measured
793: apparent speed for this component of $0.00c\pm0.04c$, compared to a
794: measured apparent speed of $-0.21c\pm0.61c$ for this same component
795: based on only the three epochs from 2000 that were analyzed in Paper~I.
796: The addition of the three new epochs thus lowers the upper limit
797: on the apparent speed of this component by a factor of ten, from
798: $0.40c$ to $0.04c$. Because of the low upper limit on the apparent
799: speed, this component most likely represents a stationary component in
800: the jet, whose apparent speed is not directly related to the bulk
801: apparent speed of the jet (see the discussion on stationary and moving
802: components in $\S$~\ref{disc}). The flux density of the component is
803: variable over the six epochs, as might be expected from a stationary
804: feature with variable jet plasma passing through it. If only the
805: three epochs from this paper are used in the fitting, then an apparent
806: speed of $0.27c\pm0.35c$ is obtained for the 15\,GHz data, and an
807: apparent speed of $0.59c\pm0.29c$ is obtained from the 22\,GHz data.
808: We expect that these differences simply represent the uncertainty
809: inherent in fitting an apparent speed from three epochs spaced over
810: only half a year, and do not represent an actual acceleration of the
811: component.
812:
813: \begin{figure*}[t]
814: \begin{center}
815: \includegraphics[scale=0.40,angle=90.0]{f10.ps}
816: \caption{Distance from the core of the center of Gaussian component C1 in PKS\,2155$-$304 as
817: a function of time. The line is the least-squares fit to outward motion with constant speed.}
818: \end{center}
819: \end{figure*}
820:
821:
822: \subsection{PKS\,2155$-$304}
823:
824: The $z=0.116$ HBL PKS\,2155$-$304 is a prototypical object of this
825: class, and has been extensively observed at high energies. In
826: particular, since 2002, its very high-energy gamma-ray flux has been
827: monitored with the H.E.S.S. array of atmospheric-Cherenkov telescopes
828: (Aharonian et al.\ 2005a). A multi-wavelength campaign undertaken in
829: 2003 is described by Aharonian et al.\ (2005b). These authors model
830: the SED of this object during a low state, and fit the spectrum with
831: several SSC model variations with Doppler boosting factors ranging
832: between 25 and 50. Of particular note is the recent dramatic flaring
833: activity at TeV energies on 28 July 2006 (Aharonian et al.\ 2007;
834: Sakamoto et al.\ 2007).
835: Bursts varying on timescales of 200 s were observed during this
836: outburst, requiring Doppler boosting factors greater than 100 if the
837: emission region has a size comparable to the Schwarzschild radius of a
838: $\sim10^{9}M_{\odot}$ black hole (Aharonian et al.\ 2007).
839:
840: Probably because of its declination, this source has not been
841: well-observed in the radio. A series of images showing the
842: large-scale radio structure of PKS\,2155$-$304 is shown by
843: Laurent-Muehleisen et al.\ (1993). The highest-resolution VLA image
844: presented by these authors shows a knot nearly 180$\arcdeg$ misaligned
845: from the VLBA jet seen in Paper~I; the lower-resolution VLA images by
846: these same authors show an extended halo of emission around the core.
847: This source has not been well-observed with VLBI, and the only other
848: published VLBI images known to us are the three-epoch set from Paper~I.
849:
850: Our new 15\,GHz VLBA images of PKS\,2155$-$304 from 2003 Sep 20 are
851: shown in Figure~8. The linear resolution of these images at the
852: distance of PKS\,2155$-$304
853: %($z=0.116$)
854: is approximately 2.1\,pc\,mas$^{-1}$. Parameters of these images are given in
855: Table~\ref{imtab}. The tick marks on the images indicate the
856: magnitude of the polarized flux and the direction of the EVPA, and the
857: color scale indicates the fractional polarization. These are the
858: first parsec-scale polarization images of this object. The
859: total-intensity morphology is consistent with that seen in Paper~I,
860: with a compact jet component to the southeast of the core (here at a
861: position angle of $\approx160\arcdeg$), and a diffuse jet with a broad
862: opening angle extending to the east. This diffuse eastern jet is more
863: easily seen in the tapered image from the same dataset shown in
864: Figure~9.
865:
866: Polarized flux is significantly detected at the location of the core
867: component in both images in Figure~8. The percentage polarization at
868: the model-fit position of the core (the origin) is 2.9\%, and the
869: electric vector position angle is $131\arcdeg$, about $30\arcdeg$
870: misaligned from the innermost jet position angle. The fractional
871: polarization increases to the west of the model-fit core position, to
872: peak values of 8--10\% (see the color scale images in Figure~8).
873: Because the observations are at only a single frequency, no
874: corrections have been made for Faraday rotation. However, the
875: measured rotation measures in BL Lac objects are typically a few
876: hundred rad m$^{-2}$ or less (Zavala \& Taylor 2003, 2004), resulting
877: in an expected error of only about 10$\arcdeg$ or so to the EVPA at
878: 15\,GHz. The foreground integrated rotation measure in this region is
879: low, resulting in a negligible correction of under 1$\arcdeg$
880: (Simard-Normandin, Kronberg, \& Button 1981).
881:
882: As in Paper~I, the visibilities are well fit by two circular
883: Gaussians, consisting of the compact core and a single jet component
884: to the southeast of the core. Parameters of the model fit are given
885: in Table~\ref{mfittab}. The core component is partially resolved with
886: a best-fit size of 0.23 mas and a lower limit to its size of 0.19 mas,
887: corresponding to an upper limit to its brightness temperature of
888: $2\times10^{10}$~K. The single jet component is slightly farther from
889: the core than the jet component seen in Paper~I, and we identify it as
890: the same component, named C1 in Paper~I. As mentioned above,
891: identifying components across a several-year gap can be risky, but the
892: structure of this source is simple, consisting of only a single jet
893: component, so we consider this identification to be
894: likely.
895: %plausible.
896: A fit
897: to the outward motion of this component using the combined four-epoch
898: dataset from this paper and Paper~I is shown in Figure~10. The fitted
899: apparent speed is $0.93c\pm0.31c$. The measurement error on the
900: apparent speed has been reduced by a factor of nearly ten over the
901: apparent speed given in Paper~I of $4.37c\pm2.88c$. The new apparent
902: speed measurement is considerably lower than the previous value, but
903: given the large associated error on the previous result, they are
904: statistically consistent at the 1.2$\sigma$ level. The component
905: faded over the three years since it was first detected, from about 50
906: mJy in early 2000 (see Paper~I) to about 15\,mJy in 2003. It has also
907: turned slightly to the south, moving from a position angle of
908: $\approx150\arcdeg$ in 2000 to $\approx160\arcdeg$ in 2003.
909:
910: \begin{table*}[t]
911: \begin{center}
912: \caption{Apparent Component Speeds in TeV HBLs}
913: \label{speedtab}
914: \begin{tabular}{l c c c c} \tableline \tableline
915: & & & & Fastest \\
916: & & Apparent Speed$^{a}$ & & Apparent Speed \\
917: Source & Comp. & (multiples of $c$) & Ref. & (multiples of $c$) \\ \tableline
918: Mkn~421 & C4 & $0.09\pm0.07$ & 1 & $0.10\pm0.02$ \\
919: & C4a & $-0.06\pm0.09$ & 1 & \\
920: & C5 & $0.10\pm0.02$ & 1 & \\
921: & C6 & $0.03\pm0.03$ & 1 & \\
922: & C7 & $0.06\pm0.01$ & 1 & \\
923: H\,1426+428 & C1 & $2.09\pm0.53$ & 4 & $2.09\pm0.53$ \\
924: Mkn~501 & C1 & $0.05\pm0.18$ & 2 & $0.54\pm0.14$ \\
925: & C2 & $0.54\pm0.14$ & 2 & \\
926: & C3 & $0.26\pm0.11$ & 2 & \\
927: & C4 & $-0.02\pm0.06$ & 2 & \\
928: 1ES\,1959+650 & C1 & $-0.11\pm0.79$ & 3 & $0.00\pm0.04$ \\
929: & C2 & $0.00\pm0.04$ & 4 & \\
930: PKS\,2155$-$304 & C1 & $0.93\pm0.31$ & 4 & $0.93\pm0.31$ \\
931: 1ES\,2344+514 & C1 & $1.15\pm0.46$ & 3 & $1.15\pm0.46$ \\
932: & C2 & $0.46\pm0.43$ & 3 & \\
933: & C3 & $-0.19\pm0.40$ & 3 & \\ \tableline
934: \end{tabular}
935: \end{center}
936: %\\ [5pt]
937: $a$: for $H_{0}=71$ km s$^{-1}$ Mpc$^{-1}$, $\Omega_{m}=0.27$, and $\Omega_{\Lambda}=0.73$.\\
938: References. --- (1) Piner \& Edwards (2005). (2) Edwards \& Piner (2002)
939: with modified cosmological parameters. (3) Piner \& Edwards (2004) (Paper I). (4) this paper.
940: \end{table*}
941:
942: \section{Discussion}
943: \label{disc}
944:
945: A major data product resulting from the high-resolution multi-epoch
946: VLBA monitoring are the apparent jet speeds. In Table~\ref{speedtab}
947: we give a table of all apparent speeds that we have measured in TeV
948: blazars over the course of our monitoring program. This table is meant
949: to update Table~3 of Paper~I, and it contains our results for Mkn~421
950: from Piner \& Edwards (2005), as well as the new apparent speeds for
951: H\,1426+428, 1ES\,1959+650, and PKS 2155$-$304 measured in this paper.
952: As noted in Paper~I, the apparent pattern speeds measured in the TeV
953: blazars are considerably slower than those measured in sources
954: selected for their compact radio emission (Kellermann et al.\ 2004), or
955: for their GeV gamma-ray emission (Jorstad et al.\ 2001). The updated
956: apparent speed measurements added in this paper strengthen the
957: statistical significance of these conclusions. Applying the
958: Kolmogorov-Smirnov test to the apparent speed distribution in
959: Table~\ref{speedtab} compared to the apparent speed distributions in
960: Kellermann et al.\ (2004) and Jorstad et al.\ (2001) shows that the TeV
961: blazars have slower pattern speeds with $>99.75\%$ confidence compared
962: to the compact radio sources of Kellermann et al.\ (2004), and with
963: $>99.99\%$ confidence compared to the EGRET blazars of Jorstad et al.
964: (2001).
965:
966: VLBA observations of the large 2 cm survey and MOJAVE survey have
967: shown that powerful jets form VLBI patterns that move at apparent
968: speeds ranging from zero up to the apparent bulk speed of the jet
969: (Lister 2006; Cohen et al.\ 2007). These authors conclude that the peak
970: apparent pattern speed measured in a jet from VLBI monitoring is then
971: a good indicator of the apparent bulk speed of the flow. In the final
972: column of Table~\ref{speedtab}, we therefore list the fastest pattern
973: speed observed for each of these TeV blazars during our monitoring.
974: All of these fastest speeds are below about 2$c$, and four of the six
975: are subluminal. One of the following two possibilities must then apply
976: to the apparent pattern speeds in the parsec-scale jets of the TeV
977: HBLs:
978: \begin{enumerate}
979: \item{The pattern speeds are related to the bulk apparent speeds as described above, and the
980: bulk apparent speeds are slower on the parsec scale in the TeV HBLs compared to the more powerful sources.}
981: \item{A much larger fraction (approaching 100\%)
982: of the HBLs have pattern speeds that are unrelated to their bulk apparent speeds
983: than do the more powerful sources.}
984: \end{enumerate}
985: Either of these two possibilities implies a fundamental difference between the parsec-scale jets of HBLs and
986: the more powerful sources, and we discuss each possibility in turn:
987:
988: \subsection{Pattern speeds related to bulk speeds}
989:
990: If the fastest pattern speeds are an indication of the bulk apparent
991: speed, as they are for the MOJAVE sources, then we have a typical bulk
992: apparent speed of about 1$c$ in the parsec-scale jets of the TeV
993: blazars (the mean and median of the final column in
994: Table~\ref{speedtab} are both about 1$c$). The bulk apparent speed is
995: a function of the bulk Lorentz factor and the viewing angle, as is the
996: Doppler factor (see equations in footnote 3). If the high Doppler
997: factors implied by the high-energy observations are combined with the
998: slow apparent speeds then very small viewing angles are obtained (many
999: under one degree), that imply unreasonable numbers of parent objects
1000: (Henri \& Saug\'{e} 2006). A statistically reasonable solution is
1001: then to adopt a lower Doppler factor (and therefore Lorentz factor) in
1002: the parsec-scale radio-emitting region than in the gamma-ray emitting
1003: region. This argument was discussed in more detail in Paper~I.
1004:
1005: A lower Doppler factor and bulk Lorentz factor for the parsec-scale
1006: radio emission in HBLs is actually suggested by many observations
1007: other than the apparent pattern speeds. Each of the observables
1008: described below depends on the bulk Doppler factor of the jet, not on
1009: a pattern speed:
1010: \begin{description}
1011: \item[a.]{The TeV blazars are significantly less variable in the radio
1012: band than are the EGRET blazars (Aller, Aller, \& Hughes 2006).}
1013: \item[b.]{The TeV blazars have low radio core brightness temperatures
1014: (see Table~\ref{mfittab}) that do not require invoking
1015: relativistic Doppler factors to reduce them below possible
1016: intrinsic limits. Typical high-power sources have much higher
1017: brightness temperatures (Kovalev et al.\ 2005; Dodson et al.\
1018: 2007).}
1019: \item[c.]{The measured radio powers of the cores suggest Lorentz
1020: factors for HBLs of around $\Gamma=3$, when this power is compared
1021: to that expected from the total power at low frequency (Giroletti
1022: et al.\ 2004a).}
1023: \item[d.]{The calculated inverse Compton Doppler factor from the VLBI
1024: cores of HBLs gives a mean lower limit to the Doppler factor of
1025: $\delta\gtrsim4$, and a mean lower limit to the Lorentz factor of
1026: $\Gamma\gtrsim2$ (see Table~7 in Giroletti et al.\ 2004a). Thus, high
1027: Lorentz factors are not required to reduce the predicted X-ray
1028: emission from the VLBI cores below the observed values.}
1029: \item[e.]{Optical and radio luminosities of HBLs could not be unified
1030: with their FR~I parent population using only transverse jet
1031: velocity structures by Chiaberge et al.\ (2000), but the required
1032: properties could be reproduced by assuming that the radio emitting
1033: region in HBLs is less beamed than the optical one, as could be
1034: expected if the jet decelerates after the higher energy emitting
1035: zone (Chiaberge et al.\ 2000).}
1036: \item[f.]{The VLBI jets of the TeV blazars assume a plume-like
1037: morphology beyond a few mas from the core, in contrast to
1038: high-power jets that may remain well-collimated to large
1039: distances. This transition to a plume-like morphology has been
1040: interpreted as evidence for entrainment induced deceleration in
1041: TeV blazars by Bicknell et al.\ (2005).}
1042: \end{description}
1043:
1044: These six observational properties all imply relatively low bulk Doppler
1045: factors and Lorentz factors in these parsec-scale jets. It is clear
1046: that radio observations consistently yield Doppler factors and Lorentz
1047: factors for the TeV blazars (and HBLs in general) about an order of
1048: magnitude lower than the high-energy observations. On the other hand,
1049: arguing for a higher Lorentz factor is the fact that no counterjets
1050: have been observed in any of the TeV HBLs. Our observations put the
1051: limit on the jet to counterjet ratio $J$ at $J\gtrsim100$ for most of
1052: the sources. The tightest limits on $J$ in a TeV blazar come from
1053: High-Sensitivity Array (HSA) observations of Mkn\,501 by Giovannini,
1054: Giroletti, \& Taylor (2007), which constrain $\beta\cos\theta>0.92$,
1055: corresponding to $\Gamma\gtrsim3$ for viewing angles of a few degrees.
1056: Thus, a parsec-scale Lorentz factor in the radio-emitting region of
1057: $\Gamma\sim3$ seems consistent with the six points given above, and is
1058: just consistent with the observed jet-to-counterjet brightness ratios.
1059: A Lorentz factor of $\Gamma\sim3$ is also consistent with the fastest
1060: apparent speeds given in Table~\ref{speedtab} (except for the very
1061: slow speeds in Mkn~421 and 1ES\,1959+650), and yields reasonable
1062: viewing angles of 2$\arcdeg$ to 8$\arcdeg$ for the apparent speeds
1063: between 0.5$c$ and 2$c$. It is not surprising that these viewing
1064: angles may be substantially less than 1/$\Gamma$, since these sources
1065: were selected based on their TeV emission which may have a narrower
1066: beaming cone.
1067:
1068: The crucial questions then become: where is the parsec-scale radio
1069: emission located relative to the site of the gamma-ray emission, and
1070: what causes its lower Lorentz factor? One possibility is that the
1071: parsec-scale radio emission (at light-year scales) is located
1072: downstream of the gamma-ray emission (at light-day scales), and that
1073: the jet decelerates as it moves out. Georganopoulos \& Kazanas (2003)
1074: considered such a jet that decelerates along its length, and showed
1075: that the required values of the Doppler factor in the inner jet are
1076: not as high as in the homogeneous models discussed above, because the
1077: fast inner portion of the jet can more efficiently upscatter
1078: blueshifted synchrotron photons from the slower outer portion. Thus,
1079: allowing an inhomogeneous model reduces some of the apparent conflict.
1080: Wang et al.\ (2004) modeled a decelerating jet where the dissipated
1081: kinetic energy of the jet is transferred to electron acceleration,
1082: they then used the resulting radiation from the accelerated electrons
1083: to successfully model the SED of several TeV blazars. Bicknell et al.
1084: (2005) have also modeled parsec-scale TeV blazar jet deceleration that
1085: they claim physically reflects a conversion of kinetic luminosity into
1086: enthalpy flux, and concluded that such deceleration can be physically
1087: consistent with the conservation of energy and momentum in this
1088: region.
1089:
1090: A second possibility is that the jet has an inhomogeneous structure
1091: transverse to the jet axis, consisting of a fast `spine' which
1092: dominates the high-energy emission, and a slower `layer', which
1093: dominates at lower frequencies, although such a structure is not
1094: mutually exclusive with the decelerating jet discussed above. For
1095: example, Ghisellini et al.\ (2005) proposed a jet with an initial {\em
1096: transverse} velocity structure consisting of a fast spine and a slow
1097: layer, and by taking into account the inverse Compton radiation each
1098: portion produced from the seed radiation coming from the other
1099: portion, they found that the resulting anisotropic inverse Compton
1100: emission from the spine may serve to decelerate it. Henri \&
1101: Saug\'{e} (2006) propose a two-flow model, consisting of a mildly
1102: relativistic outer MHD jet, and a central pair-plasma jet accelerated
1103: by the so-called `Compton rocket' effect to higher velocities of order
1104: $\Gamma\sim3$. Regardless of their origin, such transverse structures
1105: may show an observational signature in the VLBI images of jets that
1106: are resolved in the transverse direction, in the form of
1107: limb-brightening or limb-darkening, according to which region is
1108: dominating the radio emission. Observational signatures of
1109: limb-brightening have been claimed in the VLBI images of the TeV
1110: sources Mkn~501 (Giroletti et al.\ 2004b), Mkn~421 (Giroletti et al.
1111: 2006), and M~87 (Kovalev et al.\ 2007). We also found some
1112: observational evidence of limb-brightening in the Mkn~421 jet (Piner
1113: \& Edwards 2005), but it was present in only some of the transverse
1114: slices, suggesting a more complex transverse emission pattern on the
1115: scales that we investigated. The jets of the three fainter (in the
1116: radio) TeV sources considered in this paper are not detected with
1117: sufficient resolution or dynamic range in the VLBA images to discern
1118: structure transverse to the jet axis; those observations must be left
1119: to the brighter sources (or to much more sensitive images of these
1120: sources).
1121:
1122: A third possibility, that the gamma-ray emission is co-spatial with
1123: the parsec-scale jet knots, was suggested by observations of knot
1124: HST-1 in M~87 (recently reviewed by Harris et al.\ 2007). That knot,
1125: lying at a distance of $\sim100$ pc from the core of M~87, would lie
1126: at angular separations of a few mas from the core in the $z\sim0.1$
1127: TeV blazars, taking into account likely projection effects in jets
1128: with viewing angles of a few degrees. That would put features like
1129: HST-1 in M~87 co-spatial with the VLBI model components in the jets of
1130: the TeV blazars. If the high-energy emission is occurring in these
1131: components, that makes the different Lorentz factors and Doppler
1132: factors derived in the radio and the gamma-ray particularly
1133: challenging to explain.
1134:
1135: \subsection{Pattern speeds unrelated to bulk speeds}
1136:
1137: Many sources in large VLBI surveys display a mix of stationary and
1138: moving components --- but what we mean here is that even the {\em
1139: fastest} measured apparent pattern speed in a source may still be
1140: much less than the bulk apparent speed. The study of the 2 cm and
1141: MOJAVE survey apparent speeds (Cohen et al.\ 2007) found that a small
1142: fraction (about 25\%) of the powerful sources showed {\em only} slow
1143: apparent pattern speeds that were unrelated to their bulk apparent
1144: speeds based on statistical arguments (otherwise they would have
1145: unreasonably small angles to the line-of-sight, or unreasonably high
1146: intrinsic luminosities). If the bulk Lorentz factor is indeed high in
1147: the parsec-scale jets of the TeV HBLs, then the fraction of these
1148: sources that have apparent pattern speeds much less than their
1149: apparent bulk speed is closer to 100\%.
1150:
1151: Is there then some plausible physical difference in the high and
1152: low-power jets that could cause the high-power jets to form patterns
1153: that move at the bulk speed, while the low-power jets do not? A
1154: common model for particle acceleration and the formation of VLBI
1155: components is the shocked-jet model (e.g., Spada et al.\ 2001). If such
1156: rapidly moving shocks were forming and propagating to the parsec-scale in the TeV
1157: HBLs, then we might expect them to be visible in the VLBA monitoring,
1158: and they are not. Krawczynski (2007) has suggested a model where
1159: particle acceleration in the low-power sources occurs through
1160: acceleration of parallel electron-positron or electron-proton beams
1161: rather than through shock acceleration. If a source does not form
1162: strong moving shocks, then its more continuous jet may be modeled by a
1163: series of stationary Gaussians, even though the plasma is moving
1164: relativistically. Other features that are due to standing shocks, or
1165: to points where the jet bends toward the line-of-sight, may also
1166: appear in jets as stationary regions of locally-enhanced emission, as
1167: they do in high-power sources.
1168:
1169: Another possible physical difference between the high and low-power
1170: jets that may cause such a difference in pattern speeds is the opening
1171: angle of the (presumably conical) jet. This possibility is explored
1172: in the series of papers by Gopal-Krishna et al.\ (2004, 2006, 2007).
1173: According to these models, if the low-power sources have jets with
1174: considerably larger opening angles (Gopal-Krishna et al.\ 2007), then
1175: what is seen in the VLBI observations is a set of observables that
1176: have been unintentionally averaged over different viewing angles, with
1177: the small viewing-angle portion of the jet being the most Doppler
1178: boosted, and contributing the most to the weighted average. In that
1179: case, many sources may show slow apparent speeds typical of being
1180: viewed almost end-on, while the apparent speed {\em on the axis} of
1181: the jet is actually much higher. However, in all of the versions of
1182: this model that produce subluminal effective apparent speeds, the
1183: effective Doppler factor remains quite high ($\delta_{eff}\gtrsim20$,
1184: see Figures~3 and 4 of Gopal-Krishna et al.\ 2007), so that this model
1185: cannot easily explain the relatively low Doppler factors measured for
1186: the TeV blazars in the radio as discussed in points {\bf a} through
1187: {\bf f} in the previous subsection (in fact, it would predict high
1188: values for the radio variability, brightness temperature, core
1189: dominance, and inverse Compton Doppler factors).
1190:
1191: Whether or not the
1192: opening angle has a significant effect on the observables hinges on
1193: the actual value of the opening angle --- the effect is small for
1194: 1$\arcdeg$ opening angles and quite pronounced at 5$\arcdeg$ opening
1195: angles; but, perversely, measurements of blazar jet opening angles are
1196: of little use in testing the model, because the measurements of the
1197: opening angle are themselves effected by averaging over the true
1198: opening angle (see Figure~3, Gopal-Krishna et al.\ 2006). However, if
1199: the opening angles are large enough to cause a significant effect, it
1200: would seem that a substantial number of sources may be expected to
1201: have core-halo type morphology on their VLBI images (whenever the
1202: opening angle is larger than the axis viewing angle), so it would be
1203: useful to compare simulated VLBI images from this model with observed
1204: morphologies. The model also makes specific statistical predictions,
1205: e.g., that a small fraction of TeV blazars should show strongly
1206: superluminal motions (Gopal-Krishna et al. 2006), so that as the
1207: number of TeV blazars increases, it will be possible to test the
1208: observations against probability curves derived from this model.
1209:
1210: Based on all of the available observational data, we favor a
1211: combination of possibilities 1 and 2 given at the beginning of this
1212: discussion section. From all of the other arguments for a low Lorentz
1213: factor for the parsec-scale radio emission given above (points {\bf a}
1214: through {\bf f} in the previous subsection), we consider it likely
1215: that the parsec-scale radio emitting region in the TeV HBLs has a
1216: substantially lower Lorentz factor than is usually derived for the
1217: high-energy gamma-ray emitting region. A parsec-scale Lorentz factor
1218: of about 3 is indeed consistent with the peak apparent speeds of
1219: 0.5$c$ to 2$c$ in Table~\ref{speedtab}, assuming these reflect the
1220: bulk apparent speed on the axis of the jet. However, the apparent
1221: speeds in two of the TeV blazars (Mkn~421 and 1ES\,1959+650) are so low
1222: that they likely represent a sub-sample of these sources where the
1223: measured pattern speeds are much less than the bulk speed, as was also
1224: found for powerful sources by Cohen et al.\ (2007). The physical cause
1225: of this Lorentz factor difference between the high and low-frequency
1226: emitting regions (and whether any Lorentz factor gradient occurs along
1227: the jet, transverse to the jet, or both), and the physical cause of
1228: any decoupling of pattern and bulk speeds has not been definitively
1229: determined, but the sample of TeV blazars is growing rapidly, so that
1230: the observational statistics will improve markedly over the next
1231: several years.
1232:
1233: \section{Conclusions}
1234: We have presented new multi-epoch VLBI images
1235: for the three TeV blazars H~1426+428, 1ES~1959+650, and PKS~2155$-$304
1236: obtained during the years 2001 to 2004.
1237: The results for H~1426+428 are the first multi-epoch VLBI results to
1238: be presented for this source. The results
1239: for 1ES~1959+650 and PKS~2155$-$304 are combined with earlier
1240: results for these sources from Paper I to yield a longer time baseline for measurement.
1241: The major observational findings for these three sources are:
1242: \begin{enumerate}
1243: \item{H~1426+428's parsec-scale structure during this time range was well-modeled by a
1244: $\sim 17$\,mJy core and a single $\sim 3$\,mJy jet component at a position angle of
1245: approximately $-25\arcdeg$, with an apparent speed of $2.09c\pm0.53c$.}
1246: \item{1ES~1959+650 consisted of a compact core and a nearby stationary ($0.00c\pm0.04c$)
1247: jet component at a position angle of about $125\arcdeg$ and
1248: a separation of about 0.35 mas at 15 GHz. On larger scales of a few mas the
1249: jet is diffuse and directed to the north, so that this source shows an extreme apparent
1250: misalignment of about $130\arcdeg$ on parsec scales.}
1251: \item{PKS~2155$-$304 was observed with dual-circular polarization at 15 GHz.
1252: The fractional polarization at the
1253: position of the core was 3\%, and the electric vector position angle
1254: was $131\arcdeg$, about $30\arcdeg$ misaligned from the innermost
1255: jet position angle. The measured apparent speed of the single jet component
1256: was $0.93c\pm0.31c$.}
1257: \end{enumerate}
1258:
1259: We combined the new apparent speed measurements from this paper
1260: with the apparent speeds measured in TeV blazar jets from our earlier papers to form
1261: a current set of apparent speed measurements in TeV HBLs (Table~\ref{speedtab}).
1262: The mean peak apparent pattern speed in the jets of the TeV HBLs is about 1$c$.
1263: The statistical result noted in Paper I that the TeV HBLs have significantly slower
1264: apparent jet pattern speeds compared to radio-selected or GeV-selected blazars is strengthened
1265: by the new results of this paper.
1266: The Discussion section ($\S$~\ref{disc}) presented a thorough analysis of these
1267: results in the context of other radio observations and theoretical models for
1268: TeV blazar jets. Conclusions from that analysis were:
1269: \begin{enumerate}
1270: \item{The peak apparent speeds of order 0.5-2$c$ in four of the six
1271: studied TeV blazars (Table~\ref{speedtab}), when
1272: taken together with other observed radio properties discussed in
1273: $\S$~\ref{disc}, are consistent with
1274: radio jets with bulk Lorentz factors of $\Gamma\sim3$ and viewing angles of
1275: a few degrees.}
1276: \item{The very slow peak apparent speeds in two of the six studied
1277: TeV blazars in Table~\ref{speedtab} (Mkn~421 and 1ES\,1959+650) are likely pattern
1278: speeds that are unrelated to the bulk apparent speeds of the jets,
1279: which are likely to be similar to those mentioned above.}
1280: \end{enumerate}
1281:
1282: \acknowledgments
1283: We acknowledge helpful comments from the referee that improved
1284: the quality of the paper.
1285: The National Radio Astronomy Observatory is a facility of the National
1286: Science Foundation operated under cooperative agreement by Associated
1287: Universities, Inc.
1288: This research has made use of the NASA/IPAC
1289: Extragalactic Database (NED), which is operated by the Jet Propulsion
1290: Laboratory, California Institute of Technology, under contract with
1291: the National Aeronautics and Space Administration. This work was
1292: supported by the National Science Foundation under Grant Nos. 0305475
1293: and 0707523.
1294:
1295: {\it Facilities:} \facility{VLBA ()}
1296:
1297: \begin{references}
1298:
1299: \reference{}Aharonian, F., et al. 2002, A\&A, 384, L23
1300:
1301: \reference{}Aharonian, F., et al. 2003, A\&A, 403, 523
1302:
1303: \reference{}Aharonian, F., et al. 2005a, A\&A, 430, 865
1304:
1305: \reference{}Aharonian, F., et al. 2005b, A\&A, 442, 895
1306:
1307: \reference{}Aharonian, F., et al. 2007, ApJ, 664, L71
1308:
1309: \reference{}Alberdi, A., Marcaide, J.~M., Marscher, A.~P., Zhang, Y.~F., Elosegui, P., Gomez, J.~L., \&
1310: Shaffer, D.~B. 1993, ApJ, 402, 160
1311:
1312: \reference{}Albert, J., et al. 2006, ApJ, 639, 761
1313:
1314: \reference{}Aller, M.~F., Aller, H.~D., \& Hughes, P.~A. 2006, in
1315: ASP Conf. Ser. 350, Blazar Variability Workshop II: Entering the GLAST Era,
1316: ed. H.~R. Miller, K. Marshall, J.~R. Webb, \& M.~F. Aller (San Francisco: ASP), 25
1317:
1318: \reference{}Bennett, C.~L., et al. 2003, ApJS, 148, 1
1319:
1320: \reference{}Bicknell, G.~V., Safouris, V., Saripalli, L., Saxton, C.~J., Subrahamanyan, R., Sutherland, R.~S.,
1321: \& Wagner, S.~J. 2005,
1322: in AIP Conf. Proc. 745, High Energy Gamma-Ray Astronomy: 2nd International Symposium,
1323: ed. F.~A. Aharonian, H.~J. V\"{o}lk, \& D. Horns, (New York: AIP), 140
1324:
1325: \reference{}Bondi, M., Marcha, M.~J.~M., Dallacasa, D., \& Stanghellini, C. 2001, MNRAS, 325, 1109
1326:
1327: \reference{}Bondi, M., Marcha, M.~J.~M., Polatidis, A., Dallacasa, D., Stanghellini, C., \&
1328: Anton, S. 2004, MNRAS, 352, 112
1329:
1330: \reference{}B\"{o}ttcher, M. 2005, ApJ, 621, 176 (erratum 641, 1233 [2006])
1331:
1332: \reference{}Chiaberge, M.,
1333: Celotti, A., Capetti, A., \& Ghisellini, G. 2000, A\&A, 358, 104
1334:
1335: \reference{}Cohen, M.~H., Lister,
1336: M.~L., Homan, D.~C., Kadler, M., Kellermann, K.~I., Kovalev, Y.~Y., \&
1337: Vermeulen, R.~C. 2007, ApJ, 658, 232
1338:
1339: \reference{}Costamante, L., et al. 2001, A\&A, 371, 512
1340:
1341: \reference{}Costamante, L., Aharonian, F., Ghisellini, G., \& Horns, D. 2003,
1342: New Astronomy Reviews, 47, 677
1343:
1344: \reference{}Daniel, M.~K., et al. 2005, ApJ, 621, 181
1345:
1346: \reference{}Djannati-Ata{\"i}, A., et al. 2002, A\&A, 391, L25
1347:
1348: \reference{}Dodson, R., et al. 2008, ApJS, in press (arXiv:0710.5707)
1349:
1350: \reference{}Edwards, P.~G., \& Piner, B.~G. 2002, ApJ, 579, L67
1351:
1352: \reference{}Falcone, A.~D., Cui, W., \& Finley, J.~P. 2004, ApJ, 601, 165
1353:
1354: \reference{}Fossati, G., et al. 2008, ApJ, in press (arXiv:0710.4138)
1355:
1356: \reference{}Georganopoulos, M., \& Kazanas, D. 2003, ApJ, 594, L27
1357:
1358: \reference{}Ghisellini, G., Celotti, A., \& Costamante, L. 2002, A\&A, 386, 833
1359:
1360: \reference{}Ghisellini, G., Tavecchio, F., \& Chiaberge, M. 2005, A\&A, 432, 401
1361:
1362: \reference{}Giovannini, G., Giroletti, M., \& Taylor, G.~B. 2007,
1363: in ASP Conf. Ser., Extragalactic Jets: Theory and Observation from Radio to
1364: Gamma Ray, ed. T.~A. Rector \& D.~S. De Young, (San Francisco: ASP), in press
1365: (arXiv:0708.3911)
1366: %,\\ http://aftar.uaa.alaska.edu/jets2007/doc/Giovannini\_Jets2007.ppt
1367:
1368: \reference{}Giroletti, M., Giovannini, G., Taylor, G.~B., \& Falomo, R. 2004a, ApJ, 613, 752
1369:
1370: \reference{}Giroletti, M., et al. 2004b, ApJ, 600, 127
1371:
1372: \reference{}Giroletti, M., Giovannini, G., Taylor, G.~B., \& Falomo, R. 2006, ApJ, 646, 801
1373:
1374: \reference{}Gopal-Krishna, Dhurde, S., \& Wiita, P.~J. 2004, ApJ, 615, L81
1375:
1376: \reference{}Gopal-Krishna, Wiita, P.~J., \& Dhurde, S. 2006, MNRAS, 369, 1287
1377:
1378: \reference{}Gopal-Krishna, Dhurde, S., Sircar, P., \& Wiita, P.~J. 2007, MNRAS, 377, 446
1379:
1380: \reference{}Gutierrez, K., et al. 2006, ApJ, 644, 742
1381:
1382: \reference{}Harris, D.~E., Cheung, C.~C., Stawarz, L., Biretta, J.~A., Sparks, W., Perlman, E.~S., \&
1383: Wilson, A.~S. 2007,
1384: in ASP Conf. Ser., Extragalactic Jets: Theory and Observation from Radio to
1385: Gamma Ray, ed. T.~A. Rector \& D.~S. De Young, (San Francisco: ASP), in press (arXiv:0707.3124)
1386:
1387: %\reference{}Hayashida, M., et al. 2007, preprint, (astro-ph/0709.2265)
1388:
1389: \reference{}Henri, G., \& Saug\'{e}, L. 2006, ApJ, 640, 185
1390:
1391: \reference{}Horan, D., et al. 2002, ApJ, 571, 753
1392:
1393: \reference{}Jorstad, S.~G., Marscher, A.~P., Mattox, J.~R., Wehrle, A.~E., Bloom, S.~D., \& Yurchenko,
1394: A.~V. 2001, ApJS, 134, 181
1395:
1396: \reference{}Kato, T., Kusunose, M., \& Takahara, F. 2006, ApJ, 638, 653
1397:
1398: \reference{}Kellermann, K.~I., et al. 2004, ApJ, 609, 539
1399:
1400: \reference{}Kollgaard, R.~I., Gabuzda, D.~C., \& Feigelson, E.~D. 1996, ApJ, 460, 174
1401:
1402: \reference{}K\"{o}nigl, A. 1981, ApJ, 243, 700
1403:
1404: \reference{}Kovalev, Y.~Y., et al. 2005, AJ, 130, 2473
1405:
1406: \reference{}Kovalev, Y.~Y., Lister, M.~L., Homan, D.~C., \& Kellermann, K.~I. 2007,
1407: ApJ, 668, L27
1408:
1409: \reference{}Krawczynski, H., 2007, ApJ, 659, 1063
1410:
1411: \reference{}Krawczynski, H., et al. 2004, ApJ, 601, 151
1412:
1413: \reference{}Kusunose, M., \& Takahara, F. 2006, ApJ, 651, 113
1414:
1415: \reference{}Laurent-Muehleisen, S.~A., Kollgaard, R.~I., Moellenbrock, G.~A., \& Feigelson, E.~D.
1416: 1993, AJ, 106, 875
1417:
1418: \reference{}Lister, M.~L. 2006, in ASP Conf. Ser. 350, Blazar Variability Workshop II: Entering the GLAST Era,
1419: ed. H.~R. Miller, K. Marshall, J.~R. Webb, \& M.~F. Aller (San Francisco: ASP), 139
1420:
1421: \reference{}Lovell, J.~E.~J. 2000, in Astrophysical Phenomena Revealed by Space VLBI,
1422: ed. H. Hirabayashi, P.~G. Edwards, \& D.~W. Murphy (Sagamihara: ISAS), 301
1423:
1424: \reference{}Ma, C., et al.\ 1998, \aj, 116, 516
1425:
1426: %\reference{}Mutel, R.~L., Su, B., Bucciferro, R.~R., \&
1427: %Phillips, R.~B.\ 1990, \apj, 352, 81
1428:
1429: \reference{}Mutel, R.~L., \& Denn, G.~R.\ 2005, \apj, 623, 79
1430:
1431: \reference{}Petry, D., et al. 2002, ApJ, 580, 104
1432:
1433: \reference{}Piner, B.~G., Unwin, S.~C., Wehrle, A.~E., Edwards, P.~G., Fey, A.~L., \& Kingham, K.~A.
1434: 1999, ApJ, 525, 176
1435:
1436: \reference{}Piner, B.~G., \& Edwards, P.~G. 2004, ApJ, 600, 115 (Paper~I)
1437:
1438: \reference{}Piner, B.~G., \& Edwards, P.~G. 2005, ApJ, 622, 168
1439:
1440: \reference{}Piner, B.~G., Mahmud, M., Fey, A.~L., \& Gospodinova, K. 2007, AJ, 133, 2357
1441:
1442: \reference{}Rector, T.~A., Gabuzda, D.~C., \& Stocke, J.~T. 2003, AJ, 125, 1060
1443:
1444: \reference{}Sakamoto, Y., et al. 2007, ApJ, submitted
1445:
1446: \reference{}Simard-Normandin, M., Kronberg, P.~P., \& Button, S. 1981, ApJS, 45, 97
1447:
1448: \reference{}Spada, M., Ghisellini, G., Lazzati, D., \& Celotti, A. 2001, MNRAS, 325, 1559
1449:
1450: \reference{}Stirling, A.~M., et al. 2003, MNRAS, 341, 405
1451:
1452: \reference{}Urry, C.~M., \& Padovani, P. 1995, PASP, 107, 803
1453:
1454: \reference{}Wagner, R. 2007a,
1455: in ASP Conf. Ser., Extragalactic Jets: Theory and Observation from Radio to
1456: Gamma Ray, ed. T.~A. Rector \& D.~S. De Young, (San Francisco: ASP), in press
1457: (arXiv:0706.4439)
1458:
1459: \reference{}Wagner, R. 2007b, MNRAS, submitted (arXiv:0711.3025)
1460:
1461: \reference{}Wang, J., Li, H., \& Xue, L. 2004, ApJ, 617, 113
1462:
1463: \reference{}Wolter, A., Beckmann, V., Ghisellini, G., Tavecchio, F., Maraschi, L.,
1464: Costamante, L., Celotti, A., \& Ghirlanda, G. 2007,
1465: in ASP Conf. Ser., Extragalactic Jets: Theory and Observation from Radio to
1466: Gamma Ray, ed. T.~A. Rector \& D.~S. De Young, (San Francisco: ASP), in press (arXiv:0707.2735)
1467:
1468: \reference{}Zavala, R.~T., \& Taylor, G.~B. 2003, ApJ, 589, 126
1469:
1470: \reference{}Zavala, R.~T., \& Taylor, G.~B. 2004, ApJ, 612, 749
1471:
1472: \end{references}
1473:
1474: \end{document}
1475: