astro-ph0402391/ms.tex
1: %% Beginning of file 'ms.tex'
2: 
3: \documentclass[12pt,preprint]{aastex}
4: %\documentclass[manuscript]{aastex}
5: %\documentclass[preprint2]{aastex}
6: %% manuscript produces a one-column, double-spaced document:
7: %\usepackage{natbib}
8: %\usepackage{aastexug}
9: \shorttitle{Synchrotron aging in SN 1993J}
10: \shortauthors{Chandra et al.}
11: 
12: \begin{document}
13: 
14: \title{Synchrotron aging and the radio spectrum of SN 1993J}
15: 
16: \author{P. Chandra}
17: \affil{Tata Institute of Fundamental Research,
18: Mumbai 400 005, India; {\tt poonam@tifr.res.in}}
19: \affil{Joint Astronomy Programme, Indian Institute of Science,
20:     Bangalore 560 012, India}
21: 
22: \author{A. Ray}
23: \affil{Tata Institute of Fundamental Research,
24: Mumbai 400 005, India; {\tt akr@tifr.res.in}}
25: \and
26: \author{S. Bhatnagar}
27: \affil{National Radio Astronomy Observatory, Socorro, NM; 
28: {\tt sbhatnag@aoc.nrao.edu}}
29: \affil{National Centre for Radio Astrophysics,
30: Pune, 411 007, India}
31: 
32: 
33: 
34: \begin{abstract}
35: 
36: We combine the GMRT low frequency radio observations of SN 1993J
37: with the VLA high frequency radio 
38: data to get  a near simultaneous spectrum around day 3200 since explosion.
39: The low frequency measurements of the supernova determine the turnover
40: frequency and flux scale of the composite spectrum and help reveal
41: a steepening in the spectral index, $\Delta \alpha \sim 0.6$,
42: in the optically thin part of the spectrum.
43: This is the first observational evidence of a break in 
44: the radio  spectrum of a young supernova.
45: We associate this break 
46: with the phenomenon of synchrotron aging of radiating electrons.
47: From the break in the spectrum
48:  we calculate the magnetic field in the shocked 
49: region independent of the equipartition assumption between
50:  energy density of relativistic particles and magnetic energy density. 
51: We  determine the ratio of these two energy
52: densities and find that this ratio is in the range:
53: $8\times 10^{-6}-5\times 10^{-4}$.  
54: We also predict  the nature of the evolution of the 
55: synchrotron break frequency 
56: with time, with competing effects due to diffusive Fermi acceleration
57: and adiabatic expansion of the radiative electron plasma.
58: 
59: \end{abstract}
60: 
61: \keywords{supernovae: individual (SN 1993J)---radiation mechanisms: non-thermal---radio continuum: stars---magnetic fields}
62: 
63: \section{Introduction}
64: 
65: The radio spectrum of a young supernova probes the conditions in
66: the magnetized plasma where this radiation originates from relativistic 
67: electrons, which are believed to be
68: accelerated in the interface region of the supernova blast-wave
69: shock and the circumstellar medium.  The radio emission from the nearby
70: supernova SN 1993J is clearly
71: of a non-thermal nature and is argued to be due to synchrotron radiation in a
72: magnetic field amplified in the interaction region and affected by
73: synchrotron self absorption and external free-free absorption (see
74: e.g.\citet{fra98} and references therein).
75: The most critical parameter of the plasma which affects the synchrotron
76: radiation spectrum, is the strength of the magnetic field. This is often
77: estimated indirectly under assumptions of equipartition of energy between 
78: the magnetic fields and that of relativistic particles or by fitting the
79: radio flux density and turnover wavelength.
80: In many classical radio sources, such as supernova remnants (SNRs) like
81: the Crab or Cassiopeia A, or in luminous radio galaxies, the radio spectral
82: index is found to steepen at high frequencies 
83: (see e.g. \citet{kar62,eil97,mye85}). This is due to the 
84: so called synchrotron aging of the source, as during the lifetime of
85: the source, electrons with high enough energies in a homogeneous magnetic
86: field will be depleted due to efficient synchrotron radiation compared
87: with the ones with lower energies. 
88: An observation of a synchrotron break can yield a measurement 
89: of the magnetic field  {\it independent
90: of the equipartition argument} if the age of the source is known 
91: (magnetic field in the 
92: crab nebula was measured by \citet{pik56} (as quoted in \citet{shk60})
93: using this technique).
94: 
95: 
96: Multi-frequency radio studies of a supernova like SN 1993J, that is
97: bright enough in the radio bands can offer such a possibility. 
98: SN 1993J exploded on March 28, 1993. It was the archetypal type IIb 
99: supernova and provided a good opportunity to study the extragalactic 
100: supernovae in detail for being the nearest extragalactic supernova 
101: (3.6 Mpc).
102: In this paper, we discuss the near simultaneous 
103: spectrum of SN 1993J obtained by combining the GMRT low frequency data
104: with the VLA high frequency data around day 3200 since explosion.
105: We find a steepening of the spectrum by $\Delta \alpha=0.6$
106: at radio frequency of 4 GHz. We associate this break with the 
107: synchrotron cooling. Synchrotron aging  of young supernovae
108: has been discussed by \citet{fra98} and also mentioned by \citet{per02}. 
109: However we find the first clear observational signature of a 
110: spectral break in radio bands of the young supernova SN 1993J.
111: With this break frequency and the independently known age of the SN 1993J, 
112: we determine the magnetic field in the supernova. 
113: Moreover, we predict how this break frequency will
114: evolve with time, based on a quasi-static evolution of the 
115: electron spectrum under the combined effect of
116: acceleration processes, synchrotron losses and 
117: adiabatic expansion of the shell where
118: the electrons are confined.  We compare this observationally determined
119: field with the best fit magnetic field under
120: equipartition assumption and thence derive the fraction by 
121: which relativistic energy density deviates from  magnetic energy density. 
122: 
123: 
124: We briefly describe the observations of SN1993J in section 2.
125: In section 3 we combine the data with the VLA data and explore
126: spectral fits with and without synchrotron cooling breaks.
127: In section 4 we explore the cumulative effects of 
128: adiabatic expansion of the
129: supernova envelope and energy gain undergone by electrons under
130: diffusive particle acceleration upon the 
131: synchrotron cooling affected particle spectrum.
132: In section 5 we discuss the
133: evolution of the break frequency with time and the importance of wide-band
134: spectra for modeling.
135: 
136: 
137: \section{GMRT Observations}
138: 
139:  We observed SN 1993J with the Giant Meter-wave
140: Radio Telescope 
141: in 610 and 235 MHz wavebands on 2001 December 30
142: and in 1420 MHz band on 2001 October 15.
143:  We combined this
144: dataset with the high frequency VLA observations (Kindly 
145: provided by \citet{sto03})
146:  in 22.5, 14.9, 8.4, 4.8 
147: and 1.4 GHz wavebands observed on 2002 January 13 (See Table \ref{tab:1} 
148: for details). 
149: Since the supernova is 10 years old, we do not expect the flux density to
150: change by significant amount in time scale of couple of months. 
151: For example using
152: time dependence of the flux density as $F\propto t^{-0.898}$ \citep{van94}, 
153: we find that the flux density of supernova will
154: change by only 2\% from 2001 October 15 to 2001 December 30, i.e. 
155: the flux density of SN at 1.4 GHz which was 33.9 mJy on 2001 October 15
156: scaled to 33.2 mJy on 2001 December 30. The difference is well within
157: the errors quoted.
158: Flux calibrators
159: 3C48 and 3C147 were observed 
160: at the beginning and at the end of the observations. 
161: In 1420 MHz observations, we used 1034+565 as a phase calibrator, whereas,
162: 0834+555 was used as phase calibrator for 610 and 235 MHz band observations. 
163: Phase calibrator was observed for 6 minutes after every 25 
164: minutes observation on the SN. 
165: The total time spend on supernova varied from 2 hours to 4 hours.
166: The data was analyzed using Astronomical Image Processing Software (AIPS).
167: The flux density errors quoted in our GMRT datasets  are
168: $$ \sqrt{[( {\rm map\, rms})^2 + (10\%\, 
169: {\rm of\,the\,peak\,flux\,of\,SN})^2]}$$
170: The 10\% of the peak flux takes into account any possible systematic errors 
171: in GMRT as well as the variation in flux in the short time 
172: difference in the near simultaneous spectrum.
173: More details of observations, data analysis and low frequency spectra
174: of SN 1993J at other epochs will be described elsewhere \citep{cha03}.
175: 
176: \section{Modeling the composite radio spectrum}
177: 
178: When we try to fit the combined GMRT plus VLA spectrum using synchrotron 
179: self absorption (SSA) \citep{che98} or free free absorption (FFA) 
180: \citep{van94} models, we find that
181: the optically thin part of the spectrum can not be fitted with
182: a single power law. The data suggests a steepening of
183: spectrum towards higher frequency end and two power laws are required to
184: fit the optically thin part of the spectrum.
185: The normalized $\chi^2= 7.3 $ per $\,5 d.o.f.$ for 
186: SSA with a single power-law improves
187: significantly to $\chi^2=0.1$ per $\,3 d.o.f.$  
188: with SSA with  power law with a break.
189: We use SSA model to derive the magnetic field 
190: from the turn-over in the spectrum
191: \footnote{Based on the observational data for various radio supernovae, 
192: \citet{sly90} had argued that for all considered supernovae SSA is responsible 
193:  for the turn-over in the spectrum, even though other absorption
194: processes are at work.}. 
195: The best fit magnetic field at the peak of the spectrum
196: under equipartition assumption is $38.3\pm 17.1$ mG and
197: radius is $(1.81 \pm 0.07) \times 10^{17}$ cm. The break in the
198: spectrum occurs at $4.02\pm 0.19$ GHz.
199: Fig. \ref{fig1} shows the spectrum with the 
200: SSA fit with a break in the spectrum. 
201: We see that the emission spectral index $\alpha$ for frequency 
202: region below 4.02 GHz is
203: $\alpha=0.51\pm0.01$ (or $ \gamma=2.01\pm 0.02$, where $\gamma$ is the
204: particle spectral index related to $\alpha$ as, $\gamma=2 \alpha+1$). 
205: The   spectral index at 4.02 GHz towards higher frequency
206: end changes by $\Delta \alpha = 
207:  0.62 \pm 0.04$, hence emission spectral index after break towards
208: higher frequency region is $\alpha =1.13\pm 0.05$. 
209: %(i.e. corresponding 
210: %$\gamma = 2 \alpha= 2.26 \pm 0.10$). 
211: This variation in spectral index is roughly
212: consistent with that
213: predicted from the synchrotron cooling effect 
214: with continuous injection (see \citet{kar62}) and we attribute the
215: break in the spectrum of SN 1993J to synchrotron cooling.  
216: 
217: 
218: \section{Synchrotron aging and effect of other energy loss/gain processes}
219: 
220: The lifetime of the relativistic electrons undergoing synchrotron losses is 
221: given as
222: \begin{equation}
223: \tau = E/[-{(dE/dt)}_{Sync}]=1.43 \times 10^{12} B^{-3/2} {\nu}^{-1/2}\; {\rm sec}
224: \end{equation} 
225: Here we use  ${B_{\perp}}^2={(B\, {\rm Sin \theta})}^2 =(2/3)B^2$.
226: The above expression 
227: is implicitly a function of time, since the magnetic field in the
228: region of emission itself 
229: changes with time as the supernova shock moves out farther
230: into the circumstellar plasma.
231: The time variation of the synchrotron break frequency can be obtained
232: by setting: $\tau = t$, whence,
233: 
234: \begin{equation} \label{sync}
235: %{{\nu}_{break}} = {\left(\frac{t}{1.43 \times 10^{12}}\right)}^{-2} B^{-3}
236: {{\nu}_{break}} = {(t/1.43 \times 10^{12})}^{-2} B^{-3}
237: =2 \times 10^{24} B_0^{-3}\, t
238: \; {\rm Hz}
239: \end{equation}
240: 
241: Here we use $B=B_0/t$ \citep{fra98}. 
242: From the above equation 
243: (and using ${\nu}_{break}=5.12\times 10^{18} B
244:  E_{break}^2 \;{\rm Hz}$ \citep{pac69})
245: and with break frequency 4 GHz, we get magnetic 
246: field $B=0.19$ Gauss for $t=3200$ days. 
247: However this estimate of the  magnetic field 
248: does not account for 
249: other processes like diffusive Fermi acceleration
250: and adiabatic losses, likely to be important for a young
251: supernova and affecting the break frequency.
252: Diffusive Fermi acceleration
253: and adiabatic expansion processes
254: do not result in change of slope of the energy spectrum 
255: from what it was in case of pure synchrotron losses,  but
256: the frequency where the synchrotron cooling break occurs 
257: gets shifted depending upon the strength of these two 
258: competing processes \citep{kar62}.
259: Adiabatic losses shifts
260: the 'break' frequency towards lower
261: frequency with time, whereas acceleration
262: processes shift the cooling break to higher frequencies.
263: We derive below the magnetic field under cumulative 
264: effect of all these processes.
265: Since supernova is young and expanding rapidly, the adiabatic losses 
266: will be given by
267: ${dE/dt}_{Adia}=-(V/R)E=-E/t$. Here
268: $V$ is the expansion velocity, i.e. the 
269: ejecta velocity and $R$ is the radius of
270: the spherical shell. 
271: 
272: 
273: In supernovae, diffusive mechanism is assumed to be the main acceleration
274: mechanism \citep{fra98,bal92}. In this process 
275: electrons gain energy every time they cross the shock front either from
276: upstream to downstream or vice versa.
277: The average fractional momentum gain per shock crossing or recrossing is:
278: $\Delta  =(4(v_1-v_2)/3v) $ and the average time taken to
279: perform one such cycle is \citep{bal92,dru83},
280: \begin{equation}
281: t_c=\frac{4{\kappa}_{\perp}}{v}\left(\frac{1}{v_1}+\frac{1}{v_2}\right) 
282: \end{equation}
283: Here $v$
284: is the test particle velocity, $v_1$ is the upstream velocity
285: and $v_2$ is the downstream velocity, and 
286: ${\kappa}_{\perp}$ is the spatial diffusion coefficient
287: of the test particles across the ambient magnetic field, when
288: the shock front is quasi-perpendicular to the field.
289: In the rest frame of shock front,
290: $v_1=V$ and $v_2=v_1/4=V/4$ (for compression factor of 4). 
291: Hence the rate of energy gain  will be 
292: \begin{equation} \label{fermi}
293: {\left(\frac{dE}{dt}\right)}_{Fermi}=\frac{\Delta E}{t_c}=
294: \frac{E V^2}{ 20 {\kappa}_{\perp}}=\frac{E (R/t)^2}{ 20 {\kappa}_{\perp}}
295: \end{equation}
296: 
297: 
298: The break in the spectrum will occur for those
299: electron energies for which the time scales for the  
300: cumulative rate of change of electron energy
301: due to synchrotron cooling plus adiabatic losses
302: and gain through diffusive acceleration 
303: becomes comparable to the life time of the supernova \citep{kar62}.
304: Lifetime of electrons  for the cumulative energy loss rate is
305: \begin{equation}
306: \tau = \frac{E}{{(dE/dt)}_{Total}}=\frac{E}{ (R^2 t^{-2}/20 {\kappa}_{\perp})E
307: -b B^2  E^2- t^{-1}E}
308: \end{equation}
309: where first term in the denominator is synchrotron loss term with 
310: $b=1.58 \times 10^{-3}$. Setting
311: the life time $\tau=t$, break frequency can be derived as: 
312: \begin{equation} 
313: \label{synchr_freq} 
314: {\nu}_{break}=\frac{2\times 10^{24}}{{B_0}^3}
315: {\left[\frac{R^2}{20 {\kappa}_{\perp}}t^{-1/2}-2\,t^{1/2}\right]}^2
316:  \,{\rm Hz}
317: \end{equation}
318: 
319: We do not yet have an observational determination
320: of ${\kappa}_{\perp}$ for SN 1993J.
321: However in the case of SN 1987A, from the delay in the
322: switch-on of the emission between
323: 843 MHz and 4.8 GHz, this parameter was estimated to be
324: ${\kappa}_{\perp}=2\times 10^{24}$ cm$^2$s$^{-1}$
325: \citep{bal92}. This  
326: is relatively independent of the density and magnetic field of the two
327: clumps in the CSM into which the SN 1987A shock was running
328: into in the first 1500-1600 days. 
329: We use a slightly different $\kappa_{\perp}$ for SN 1993J, since
330: there is evidence that the compression ratio $\rho$ of gas or plasma across
331: the shock in SN 1993J is higher than that of SN 1987A
332: \footnote{Diffusive acceleration predicts a flattening of the 
333: spectrum at a spectral index $\alpha = 3/[2(\rho -1)]$; 
334: which gives 
335: $\rho =4$ in case of SN 1993J in contrast to $\rho = 2.7$ for SN 1987A 
336: \citep{bal92}.}.
337: Therefore, we estimate that ${\kappa}_{\perp}$ for SN 1993J is
338:  (scaled by the higher compression ratio in SN 1993J):
339: ${\kappa}_{\perp}=\frac{4}{2.7}\times
340: 2 \times 10^{24}$ cm$^2$s$^{-1}$ i.e. $2.96 \times 10^{24}$ cm$^2$s$^{-1}$.
341: From VLBI observations \citep{bar02}, 
342: the extrapolated angular radius of SN 1993J
343: on day 3200 is $\sim 5012\, {\rm \mu as}$ i.e. $\sim 2.65\times 10^{17}$ cm. 
344: Using ${\kappa}_{\perp}$ and R in Eq. \ref{synchr_freq} we find magnetic field
345: $B=0.33\pm0.01 $ Gauss, from the observationally determined break.
346: 
347: 
348: From our SSA best fit to the radio spectrum, 
349: we have a value of the best fit 
350: equipartition magnetic field $B_{eq} = 38.2\pm17.1$ mG.
351: If $a$ is the ratio between relativistic electron energy density, 
352: $ U_{rel}$,
353: to magnetic energy density, $ U_{mag}$, ($a=U_{rel}/U_{mag}$), 
354: then magnetic field $B$ depends on equipartition
355: factor $a$ as $B \propto {a}^{-4/(2 \gamma +13)}$ (see \citet{che98}).
356: Hence, we get equipartition fraction $a$ i.e.  
357: $U_{rel}/U_{mag} =  8.5 \times 10^{-6}-5.0 \times10^{-4}$. 
358: This small value
359: of the ratio of particle energy density vs magnetic energy density indicates
360: that the plasma kinetics is dominated by magnetic field and associated
361: turbulence.
362:                                                                                 
363: \section{Discussion and Conclusions}
364: 
365: We see in the last section that the magnetic field
366: calculated from the break in the spectral index
367: is $0.33 $ Gauss, which is
368: $\sim$1.4 times higher than that expected from an extrapolation of 
369: \citet{fra98} ($B=0.24$ Gauss at day 3200). If we took only the
370: synchrotron cooling effect and neglected adiabatic expansion and
371: diffusive Fermi acceleration, we get magnetic field $B=0.19$ Gauss,
372: which is in closer agreement with that of \citet{fra98}.
373: However, at this young age of the supernova the effects of
374: adiabatic expansion and diffusive Fermi acceleration are likely
375: to be significant as seen in SN1987A \citep{bal92} and
376: hence these effects should not be neglected.
377: 
378: One can estimate the importance of the diffusive acceleration term, if 
379: one is able to follow how the break frequency evolves with time.
380: In Eq. \ref{synchr_freq}, first term is the contribution of the acceleration
381: and second term is the contribution of adiabatic expansion and synchrotron
382: losses.  We note that for the estimated value of 
383: ${\kappa}_{\perp}$ at the present epoch, diffusive Fermi acceleration 
384: dominates and will continue to dominate over the 
385: adiabatic losses until about 20 years since explosion.
386: Therefore, at present epoch the break frequency evolves as 
387: ${\nu}_{break} \propto t^{-1}$. 
388: After 20 years when the acceleration will cease 
389: to dominate over adiabatic expansion and  
390: the break frequency will increase as: 
391: ${\nu}_{break} \propto t$. 
392: 
393: Since we do not have an independent method to 
394: estimate the spatial diffusion coefficient 
395: ${\kappa}_{\perp}$ for SN 1993J, we have used the linearly scaled
396:  (by the respective  compression 
397: ratios, $\rho$) value of ${\kappa}_{\perp}$ measured for
398: SN 1987A from direct radio observations (see \citet{bal92}). 
399: However, we can directly calculate the 
400: value of  ${\kappa}_{\perp}$ from the (measurable) rate
401: of change of ${\nu}_{break}$, i.e. from the expression:
402: \begin{equation}
403: \frac{d {\nu}_{break}}{dt} = \frac{2 \times 10^{24}}{ B_0^{3}} 
404: \left[\frac{R^2 }
405: {20 {\kappa}_{\perp}}t^{-1/2}\,-2\,t^{1/2}\right]\left[-\frac{R^2 }
406: {20 {\kappa}_{\perp}}t^{-3/2}\,-2\,t^{-1/2}\right] {\rm Hz/sec}
407: \end{equation}
408: For the present epoch and with the estimated parameters as above, we calculate 
409: that the break frequency is changing at the rate of 1.2 GHz/year.
410: Thus a few more multi-frequency spectral observations across GMRT and
411: VLA bands, 
412: separated by a few years, will observationally determine the temporal
413: variation in the break frequency 
414: and underline the importance of the diffusive acceleration effects.
415:      
416:                                                                                 
417: The combination of multi-frequency
418: radio spectrum across GMRT and VLA bands
419: is critical for deriving the above results.
420: In Fig. \ref{fig2} we show a comparison of synchrotron self absorption
421: model (with a single optically-thin power-law index) fitted 
422: only to the low frequency data (0.22 GHz to 1.4 GHz) versus such a model
423: fit obtained with only the higher frequency data (1.4 GHz to 22.5 GHz).
424: This comparison shows that while the model fitted only to 
425: the low frequencies over-predicts
426: the flux density at high frequencies, the model
427: fitted only to high frequencies on the other hand fails 
428: to account for both synchrotron cooling break and 
429:  seriously under-predicts the low frequency flux densities.
430: The comparison underscores the importance of broad band observations
431: for determining the physical processes taking place in the supernova.
432: 
433: 
434: 
435: We note the uncertainity in 
436: %the spatial diffusion coefficient 
437: ${\kappa}_{\perp}$
438: for SN 1993J as also emphasised by the referee who suggests that
439: ${\kappa}_{\perp}$ may not be a constant, and points out a paper 
440: due to \citet{rey98}. This paper {\it assumes} that 
441: ${\kappa}_{\perp}$ is proportional to the particle energy. This dependence may
442: affect the determination of the magnetic
443: field (see Eq. \ref{synchr_freq}). Only future observations of the
444: break frequency evolution
445: with time will directly put constraints on ${\kappa}_{\perp}$ for SN 1993J.
446: We also note that our results are based on the assumption
447: that the acceleration and synchrotron losses are taking place in
448: the same region. However, if the regions of the two processes are not
449: substantially overlapping, the synchrotron break frequency will not be affected
450: by acceleration. Even in that case, the magnetic field is
451: much higher than the equipartiton
452: magnetic field and the plasma is still dominated by the magnetic
453: energy density.
454: 
455: 
456: \acknowledgements 
457: 
458: We thank Kurt Weiler and Christopher Stockdale
459: for providing us the high frequency VLA data. 
460: We thank the NRAO staff for providing AIPS.
461: We thank the staff of the GMRT that made these observations possible. 
462: GMRT is run by the National Centre for Radio Astrophysics of the 
463: Tata Institute of Fundamental Research. We thank the anonymous referee
464: for useful comments.
465: % All of this work was done using 
466: %computers running the GNU/Linux operating system and we thank all the 
467: %numerous contributors to this software.
468: 
469: \begin{thebibliography}{} 
470: \bibitem[Ball \& Kirk(1992)]{bal92} Ball, L., Kirk, J. G., 1992
471: \apj, 396, L39 
472: \bibitem[Bartel et al.(2002)]{bar02} Bartel, N., Bietenholz,M.F., Rupen, M.P., 
473: et al 2002 \apj, 581, 404
474: \bibitem[Chandra et al.(2003)]{cha03} Chandra, P., Ray, A., 
475: Bhatnagar, S. 2003 submitted to \apj 
476: \bibitem[Chevalier(1998)]{che98} Chevalier, R. 1998 \apj, 499, 810 
477: \bibitem[Drury(1983)]{dru83} Drury, L.O'C. 1983 Rep. Prog. Phys., 46, 973
478: \bibitem[Eilek et al.(1997)]{eil97} Eilek, J.A., Melrose, D.B., 
479: Walker, M.A., 1997 \apj, 483, 282 
480: \bibitem[Fransson \& Bjornsson(1998)]{fra98} Fransson, C., \& Bjornsson, C. 
481: 1998 \apj, 509, 861
482: %\bibitem[Ginzburg \& Syrovatskii(1968)]{gin68} 
483: %Ginzburg, V., \& Syrovatskii, S. 
484: %1968 AP \& SS, 1, 442
485: %\bibitem[Kaplan(1956)]{kap56} Kaplan, S.A., 1956, Sov. Phys., 2, 203
486: \bibitem[Kardashev(1962)]{kar62} Kardashev, N.S., 1962, Sov. Astr., 6, 317
487: %\bibitem[Longair(1994)]{lon94} Longair, M.G., 1994 High Energy Astrophysics,
488: %Vol 2 
489: \bibitem[Myers \& Spangler(1985)]{mye85} Myers, S.T., Spangler, S.R., 1985
490: \apj, 291, 52
491: \bibitem[Pacholczyk(1969)]{pac69} Pacholczyk, A.G., 1969 Radio 
492: Astrophysics 
493: \bibitem[Perez-Torres et al.(2002)]{per02} Perez-Torres, M.A., Alberdi, A., 
494: Marcaide, J.M. 2002 A\&A, 394, 71 
495: \bibitem[Pikelner(1956)]{pik56} Pikelner, S.B. 1956, Astr. Zhur. SSSR, 33, 785
496: \bibitem[Reynolds(1998)]{rey98} Reynolds, S. P. 1998, \apj, 493, 375
497: \bibitem[Shklovskii(1960)]{shk60} Shklovsky, I. S. 1960, Cosmic Radio Waves
498: (Harvard Univ Press) 
499: %\bibitem[Siah \& Wiita(1990)]{sia90} Siah, M.J., Wiita, P.J. 1990, \apj, 
500: %363, 411
501: %\bibitem[Scott \& Readhead(1977)]{sco77} Scott, M.A., Readhead, A.C.S. 
502: %1977 \mnras, 180, 539
503: \bibitem[Slysh(1990)]{sly90} Slysh, V.I. 1990, Sov. Astron. Lett., 16(5), 339
504: \bibitem[Stockdate et al. (2003)]{sto03} Stockdale, C., Weiler, K., VanDyk, S., Sramek, R., Panagia, N., Private Communication
505: \bibitem[van Dyk et al.(1994)]{van94} van Dyk, S. D., Weiler, K., Sramek, R., et
506: al. 1994 ApJL, 432, 115
507: \end{thebibliography}     
508: 
509: \clearpage
510: 
511: 
512: \begin{table}
513: %\scriptsize
514: \caption{Observations of the spectrum of SN 1993J
515: \label{tab:1}}
516: \begin{tabular}{ccccc}
517: \tableline\tableline
518: Date of  & Days since & Frequency &  Flux density & rms\\
519: Observation&  explosion & in GHz  & mJy   & mJy \\
520: \tableline\tableline
521: Dec 31,01                  & 3199 & 0.239  & 57.8 $\pm$7.6   & 2.5\\
522: Dec 30,01                  & 3198 & 0.619  & 47.8 $\pm$5.5   & 1.9\\
523: Oct 15,01                  & 3123 & 1.396  & 33.9 $\pm$3.5   & 0.3\\
524: Jan 13,02\tablenotemark{a} & 3212 & 1.465  & 31.44$\pm$4.28  & 2.9\\
525: Jan 13,02\tablenotemark{a} & 3212 & 4.885  & 15   $\pm$0.77  & 0.19\\
526: Jan 13,02\tablenotemark{a} & 3212 & 8.44   & 7.88 $\pm$0.46  & 0.24\\
527: Jan 13,02\tablenotemark{a} & 3212 & 14.965 & 4.49 $\pm$0.48  & 0.34\\
528: Jan 13,02\tablenotemark{a} & 3212 & 22.485 & 2.50 $\pm$0.28  & 0.13\\
529: \tableline\tableline
530: \end{tabular}
531: \tablenotetext{a} {VLA data points, \citep{sto03}}
532: \end{table}
533: 
534: \clearpage
535: 
536: \begin{figure}
537: \plotone{f1.eps}
538: \caption{ The spectrum of SN 1993J with synchrotron self-absorption
539: with a break in the spectral index at high frequencies. 
540: At low frequencies, the spectral emission index $\alpha$ is 0.51 
541: before break and after the
542: break, in high frequency regime, it is 1.13. \label{fig1}}
543: \end{figure}
544: 
545: \clearpage                                                                                
546: \begin{figure}
547: \plotone{f2.eps}
548: %\plottwo{f2a.eps}{f2b.eps}
549: \caption{Wide band spectrum of SN 1993J. Synchrotron self absorption 
550: fit to "only" low frequency data
551: (solid line) and "only" high frequency data (dashed line).\label{fig2}}
552: \end{figure}
553: 
554: \end{document}
555: %% End of file `ms.tex'.
556: