1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass[manuscript]{aastex}
3: %\documentclass[preprint2]{aastex}
4: %\documentclass{aastex}
5: %\usepackage{emulateapj5}
6:
7: \def\xr {X--ray}
8: \newcommand{\gcc}{g~cm$^{-3}\ $}
9: \newcommand{\sfun}[2]{$#1(#2)\ $}
10: \newcommand{\rhonot}{$\rho_{\circ}\ $}
11: \newcommand{\msun}{$M_{\odot}\ $}
12: \newcommand{\greq}{$\stackrel{>}{ _{\sim}}$}
13: \newcommand{\lteq}{$\stackrel{<}{ _{\sim}}$}
14: \newcommand{\etal}{{\it et al.}}
15: \newcommand{\lsim}{\raisebox{-0.3ex}{\mbox{$\stackrel{<}{_\sim} \,$}}}
16: \newcommand{\gsim}{\raisebox{-0.3ex}{\mbox{$\stackrel{>}{_\sim} \,$}}}
17: \def\gta{\ifmmode {\mathbin{\lower 3pt\hbox %> or of order
18: {$\,\rlap{\raise 5pt\hbox{$\char'076$}}\mathchar"7218\,$}}}
19: \else {${\mathbin{\lower 3pt\hbox
20: {$\rlap{\raise 5pt\hbox{$\char'076$}}\mathchar"7218\,$}}}
21: $}\fi}
22: \def\lta{\ifmmode {\,\mathbin{\lower 3pt\hbox %< or of order
23: {$\,\rlap{\raise 5pt\hbox{$\char'074$}}\mathchar"7218\,$}}}
24: \else {${\mathbin{\lower 3pt\hbox
25: {$\rlap{\raise 5pt\hbox{$\char'074$}}\mathchar"7218\,$}}}
26: $}\fi}
27:
28:
29: \shorttitle {BURST RISE OSCILLATIONS FROM 4U 1636--536}
30: \shortauthors {Bhattacharyya and Strohmayer}
31:
32: \begin{document}
33:
34: \title {Evidence for Harmonic Content and Frequency Evolution of
35: Oscillations during the Rising Phase of X-ray Bursts from 4U
36: 1636--536}
37:
38: \author {Sudip Bhattacharyya\altaffilmark{1,2}, and Tod
39: E. Strohmayer\altaffilmark{2}}
40:
41: \altaffiltext{1}{Department of Astronomy, University of Maryland at
42: College Park, College Park, MD 20742-2421}
43:
44: \altaffiltext{2}{X-ray Astrophysics Lab,
45: Exploration of the Universe Division,
46: NASA's Goddard Space Flight Center,
47: Greenbelt, MD 20771; sudip@milkyway.gsfc.nasa.gov,
48: stroh@clarence.gsfc.nasa.gov}
49:
50: \begin{abstract}
51:
52: We report on a study of the evolution of burst oscillation properties
53: during the rising phase of X-ray bursts from 4U 1636--536 observed
54: with the proportional counter array (PCA) onboard the Rossi X-Ray
55: Timing Explorer (RXTE). We present evidence for significant harmonic
56: structure of burst oscillation pulses during the early rising phases
57: of bursts. This is the first such detection in burst rise
58: oscillations, and has interesting implications for constraining
59: neutron star structure parameters and the equation of state models of
60: matter at the core of a neutron star. The detection of harmonic
61: content only during the initial portions of the burst rise appears
62: consistent with the theoretical expectation that with time the
63: thermonuclear burning region becomes larger, and hence the fundamental
64: and harmonic amplitudes both diminish. We also find, for the first
65: time from this source, strong evidence of frequency increases during
66: burst rise. The timing behavior of harmonic content, amplitude, and
67: frequency of burst rise oscillations may be important in understanding
68: the spreading of thermonuclear flames under the extreme physical
69: conditions on neutron star surfaces.
70:
71: \end{abstract}
72:
73: \keywords{equation of state --- methods: data analysis --- stars:
74: neutron --- X-rays: binaries --- X-rays: bursts --- X-rays: individual
75: (4U 1636--536)}
76:
77: \section {Introduction} \label{sec: 1}
78:
79: Millisecond period brightness oscillations, ``burst oscillations'',
80: during thermonuclear (type I) X-ray bursts from the surfaces of
81: neutron stars in low mass X-ray binary (LMXB) systems result from an
82: asymmetric brightness pattern on the rotating stellar surface
83: (Chakrabarty et al. 2003; Strohmayer, \& Bildsten 2003). This timing
84: feature reveals the stellar spin period and can provide important
85: information about the other stellar parameters (radius, mass, etc.;
86: Miller, \& Lamb 1998; Nath, Strohmayer, \& Swank 2002; Muno, \"Ozel,
87: \& Chakrabarty 2002). Theoretical modelling of these oscillations has
88: the potential to constrain the equation of state (EOS) of the dense
89: matter at the core of a neutron star (see, for example Bhattacharyya
90: et al. 2005). This can be most effectively done if the burst
91: oscillation has some harmonic content, as the fitting of a pure
92: sinusoidal lightcurve does not put strong constraints on the stellar
93: parameters. Although burst oscillations have been discovered from more
94: than a dozen LMXBs, only one source (the accreting millisecond pulsar
95: XTE J1814-338) has shown significant harmonic content (during burst
96: decay; see Strohmayer et al. 2003; Watts, Strohmayer, \& Markwardt
97: 2005). In this Letter we focus on oscillations during burst rise,
98: because at the beginning of the burst, the size of the burning region
99: is theoretically expected to be at its smallest, and the harmonic
100: content should theoretically be larger, and hence might be detected.
101:
102: The study of burst oscillations during burst rise is important for
103: another reason. At the onset of the burst, a small portion of the
104: stellar surface is ignited, and then the burning region spreads and
105: engulfs the whole stellar surface during the burst rise (Spitkovsky,
106: Levin, \& Ushomirsky 2002). This is natural, because simultaneous
107: ignition of the whole stellar surface would require very fine tuning.
108: Understanding this spreading is important, as it involves complex
109: nuclear physics and geophysical fluid dynamics (with significant
110: stellar spin), under conditions of extreme gravity, magnetic field and
111: radiation pressure. Other than the burst rise lightcurve, the time
112: evolution of three properties (frequency, amplitude, and harmonic
113: content) of burst oscillations are the most useful observational tools
114: to investigate this problem. Studies to date have not found harmonic
115: content during burst rise. Strohmayer, Zhang, \& Swank (1997) found
116: evidence for a decrease in amplitude during burst rise from an
117: analysis of RXTE data from the LMXB 4U 1728-34, and frequency increase
118: during the burst rise from the accreting millisecond pulsar SAX
119: J1808.4--3658 has also been reported (Chakrabarty et al. 2003). In
120: this Letter, we report the finding of frequency evolution during the
121: rise of several bursts, as well as evidence for significant harmonic
122: content of burst oscillations at the beginning of burst rise.
123:
124: \section {Data Analysis and results} \label{sec: 2}
125:
126: More than a hundred bursts have so far been observed from the LMXB 4U
127: 1636--536 by the proportional counter array (PCA) on board RXTE. By
128: analysing the archival data we find that 23 of them show at least
129: moderately strong oscillations during burst rise. In order to search
130: sensitively for harmonic content, it is necessary to maximize the
131: fundamental power by taking any frequency evolution into account (see
132: Miller 1999; Strohmayer \& Markwardt 1999; Muno et
133: al. 2002). Therefore, first we explore frequency evolution during
134: burst rise using the following procedures: (1) we calculate dynamic
135: power spectra (Strohmayer \& Markwardt 1999) using a time duration
136: that is small enough to resolve the burst rise, but is still large
137: enough to accumulate significant signal power. The resulting dynamic
138: spectra (panel {\it a} of Fig. 1, and Fig. 2) provide an indication of
139: the frequency evolution behavior. (2) We carry out a phase timing
140: analysis (Muno et al. 2000) to confirm the indications in the dynamic
141: spectra. We divide the burst rise time interval into several bins of
142: a fixed chosen length, and then assuming a frequency evolution model,
143: we calculate the average phase $(\psi_k)$ in each bin $(k)$. The
144: corresponding $\chi^2$ is calculated using the formula $\chi^2 =
145: \sum^M_{k=1} (\psi_k-\bar\psi_k)^2/\sigma^2_{\psi_k}$ (Strohmayer \&
146: Markwardt 2002), where $M$ is the number of bins, and $\sigma_{\psi_k}
147: = 1/\surd(Z_1^2)$ $(Z_1^2$ is the fundamental power in the
148: corresponding bin$)$. We find the best fit parameter values for a
149: particular frequency evolution model by minimizing this $\chi^2$ and
150: we calculate the uncertainty in each parameter by increasing the
151: $\chi^2$ value by the appropriate amount (Press et al. 1992). (3) To
152: confirm these results, we calculate the total fundamental power
153: $(Z_1^2)$ during the burst rise time interval using these best fit
154: frequency evolution model parameter values, and ensure that this power
155: is close to the maximum power obtained from any parameter values.
156:
157: We found four bursts with significant frequency evolution during the
158: rise. These are listed in Table 1. For burst A (panel {\it a},
159: Fig. 1), a constant frequency model gives a reduced $\chi^2 =
160: 195.03/6$ and fundamental power $Z_1^2 = 57.41$. We also calculated
161: the phase residuals (see Strohmayer \& Markwardt 2002 for details) for
162: this model. The large systematic deviations from the mean value
163: (panel {\it b}, Fig. 1), and the very high $\chi^2$ indicate that a
164: constant frequency model for burst A can be strongly rejected. Next,
165: we add a linear term to the frequency model, and find that the
166: corresponding reduced $\chi^2 = 33.76/5$, and $Z_1^2 = 134.15$. This
167: is a better fit, but still not statistically acceptable. Therefore, we
168: include a nonlinear term to the frequency model. This model has a
169: reduced $\chi^2 = 2.00/4$, and $Z_1^2 = 170.42$, and the corresponding
170: phase residuals have small random deviations from the mean value
171: (panel {\it c}, Fig. 1). Hence, we conclude, that for burst A, the
172: data strongly indicate a nonlinear frequency increase (Table 1). We
173: fit constant frequency models to the other bursts listed in Table
174: 1. For bursts B, C, \& D, the reduced $\chi^2$ $(Z_1^2)$ values for
175: this model are 225.58/6 (86.91), 43.84/6 (86.16), \& 142.11/5 (20.42)
176: respectively. These are all poor fits, and hence the constant
177: frequency model for these bursts can also be strongly rejected. Next,
178: we include a linear term in the frequency models for these bursts.
179: The corresponding reduced $\chi^2$ $(Z_1^2)$ values are 7.16/5
180: (200.35), 12.70/5 (173.18), \& 5.11/4 (89.75) respectively. These fits
181: are acceptable for bursts B \& D. For burst C, the high power value
182: and the good visual fit of this frequency model to the power contours
183: (panel {\it b}, Fig. 2) suggest that this model is on average correct,
184: and the high reduced $\chi^2$ value may be caused by fluctuation of
185: the frequency on short time scales. Therefore, for bursts B, C, \& D,
186: a linear frequency increase (Table 1 \& Fig. 2) is acceptable. In
187: panel {\it d} of Fig. 1, we show the rms amplitude variation with time
188: during the rise of burst A. The fundamental amplitude shows an
189: initially fast and then a slower decrease. Moreover, there is some
190: weak indication of a significant 1st harmonic amplitude during the
191: first $\sim 1/2$ second of this burst.
192:
193: Using our best fitting frequency evolution models, we next searched
194: for harmonic content in the burst oscillations. Individually, none of
195: the bursts shows strong harmonic power, therefore, we added the bursts
196: coherently to get more signal. For this purpose, we chose the bursts
197: with fundamental $Z^2$ power $> 30$ (for the constant frequency model)
198: during burst rise, and added them (nine in number; listed in Table 2)
199: coherently (i.e., we found the constant phase offset for each burst
200: which maximized the total fundamental power). The total co-added,
201: phase-folded lightcurve does not give a significant harmonic power. A
202: possible reason for this may be that during a significant portion of
203: the burst rise interval, the size of the burning region is large
204: enough that the harmonic amplitude is too small to be detected. To
205: address this possibility we consider five time intervals (starting at
206: the time of burst onset) of length; 1/4th, 1/3rd, half, 2/3rd, and all
207: of the rise time. For each interval fraction, and for each burst, we
208: consider that fraction of the total rise time, fit it with our
209: frequency evolution model to get best fit parameter values, and use
210: these best fit values to calculate the phases. We do this for all the
211: nine bursts and then add them coherently (separately for each interval
212: fraction). We find that the interval comprising 1/3rd of the rise
213: time gives the highest harmonic power $(17.52)$. The search at the
214: harmonic frequency in any independent subinterval is essentially a
215: single trial search (see Miller 1999), so we can estimate the
216: significance of this value using a $\chi^2$ distribution with 2
217: degrees of freedom. This gives a single-trial probability of
218: $1.57\times 10^{-4}$ to find a power as high by chance. We searched 5
219: intervals, but they are not all independent, so the number of trials
220: is between 1 and 5. Using 5 to get a conservative, lower bound on the
221: probability gives $7.85 \times 10^{-4}$, which is still a bit better
222: than a $3\sigma$ detection. As we increase the time interval, (that
223: is, use more of the rise), the harmonic power decreases gradually,
224: which is consistent with the expectation that the burning region
225: becomes larger, and hence the harmonic content diminishes. Due to the
226: consistency with this theoretical expectation, and the better than
227: $3\sigma$ harmonic power for 1/3rd of the rise interval, we conclude
228: that harmonic content in the burst rise oscillations has been
229: marginally detected.
230:
231: We extracted the corresponding (i.e., for 1/3rd of the rise time
232: interval) phase-folded lightcurve from the data (after subtraction of
233: the persistent emission) of the nine bursts. We fit it with two
234: models: (1) a single sinusoid (the fundamental) around a constant
235: level, and (2) two sinusoids (fundamental and 1st harmonic) around a
236: constant level. The former one gives a reduced $\chi^2$ value
237: $26.26/13$, while this value for the latter model is $8.33/11$ (Table
238: 2). These results support our finding that the addition of a 1st
239: harmonic provides a better description of the data. Fig. 3 shows the
240: data, the best fit model (model 2 of Table 2), and all the components
241: of the model.
242:
243: \section {Discussion}
244:
245: In this Letter, we report two new observational results: (1) the first
246: detection of frequency evolution (increase) of burst rise oscillations
247: from 4U 1636--536, and (2) the first evidence for harmonic content of
248: burst rise oscillations (from any source). These effects may be a
249: direct result of thermonuclear flame propagation on the stellar
250: surface (Bhattacharyya \& Strohmayer 2005a; 2005b). For example,
251: consider ignition of a burst off of the equator in the northern
252: hemisphere. Initially the small burning region (Spitkovsky et
253: al. 2002) can produce both a large fundamental and harmonic amplitude,
254: more or less consistent with the observations. The frequency, when
255: first observed, is at its lowest value and then increases
256: monotonically. At least two effects can account for this behavior;
257: hydrostatic expansion---and subsequent spin-down---makes the burning
258: region slip westward (on a star rotating eastward; see Strohmayer,
259: Jahoda, Giles, \& Lee 1997; Cumming \& Bildsten 2001; Cumming et
260: al. 2002; Spitkovsky et al. 2002; Bhattacharyya \& Strohmayer 2005b),
261: and the southbound front will slip even further westward due to
262: conservation of angular momentum (Bhattacharyya \& Strohmayer 2005b).
263: Thus, the hot region initially has a net retrograde drift in the frame
264: of the neutron star, so the observed frequency is less than the spin
265: frequency. As the front approaches the equator it spreads faster,
266: eventually forming a more or less symmetric equatorial belt
267: (Spitkovsky et al. 2002). This can plausibly reduce the pulsation
268: amplitude in both the fundamental and harmonic, and the westward drift
269: slows because mass elements moving southward below the equator now
270: drift eastward, conserving angular momentum. Thus, the frequency
271: increases from its initial (low) value. Once the equatorial belt has
272: been ignited, it seems likely that residual asymmetry associated with
273: the initial, northbound burning front is responsible for the observed,
274: lower amplitude oscillations. Detailed calculations of the flux from
275: such a spreading burning front are required to explore this scenario
276: quantitatively, but the discussion above provides a plausible
277: qualitative understanding of many of the observed properties (more
278: detailed discussions are in Bhattacharyya \& Strohmayer 2005b).
279:
280: As noted above, the harmonic content found at the beginning of the
281: bursts is consistent with the expected size of the burning region. We
282: have computed theoretical models to show that the inferred harmonic
283: content is theoretically possible. For example, the harmonic to
284: fundamental amplitude ratio $(a_{\rm 2}/a_{\rm 1})$, and the relative
285: phase difference $(\epsilon_{\rm 1} - \epsilon_{\rm 2})$ of the
286: components can be reproduced reasonably well with a model assuming
287: emission from a circular hot spot (Bhattacharyya et al. 2005). Using a
288: dimensionless stellar radius-to-mass ratio $Rc^2/GM = 4.5$, stellar
289: mass $M = 1.5 M_\odot$, observer's inclination angle $i = 70^{\rm o}$,
290: $\theta$-position of the center of the circular burning region
291: $\theta_{\rm c} = 50^{\rm o}$, angular radius of the burning region
292: $\Delta\theta = 25^{\rm o}$, beaming parameter $n = 1.0$, and
293: blackbody temperature of the burning region $T_{\rm BB} = 2.0$~keV, we
294: can explain the relative strength and phasing of the fundamental and
295: harmonic components. Here, we have assumed a Schwarzschild spacetime,
296: and the beaming parameter $n$ gives a measure of the beaming in the
297: frame corotating with the star (Bhattacharyya et al. 2005). These
298: model parameter values give $a_{\rm 2}/a_{\rm 1} = 0.24$ and
299: $\epsilon_{\rm 1} - \epsilon_{\rm 2} = 0.41$, while for the data
300: (Fig. 3), $a_{\rm 2}/a_{\rm 1} = 0.24\pm0.06$ and $\epsilon_{\rm 1} -
301: \epsilon_{\rm 2} = 0.36\pm0.02$. The constant level, required by the
302: observed lightcurve, can plausibly be supplied by a symmetric
303: belt-like component of the burning region (as mentioned earlier).
304: We note that waves in the surface layers of a neutron star (Heyl
305: 2005; Lee \& Strohmayer 2005) are unlikely to produce a significant
306: harmonic content, and hence at least for the rising phases of these
307: bursts, a wave interpretation of oscillations seems disfavored.
308:
309: Our results suggest that detailed studies of the onset of bursts
310: can, in principle, provide insight into the structure of neutron
311: stars, and how thermonuclear flames propagate on their surfaces.
312: Moreover, the studies of bursts with oscillations during both rise and
313: tail can determine whether these two oscillations are phase-connected,
314: and hence of the same origin. Unfortunately, current studies are
315: observationally limited by the detected count rates. However, a way
316: out of this is to add a number of bursts to increase the signal to
317: noise ratio, as we have done in this Letter to detect the harmonic
318: content.
319:
320:
321: \acknowledgments
322:
323: This work was supported in part by NASA Guest Investigator grants.
324:
325: \newpage
326: \begin{thebibliography}{}
327:
328: \bibitem[]{343} Bhattacharyya, S., \& Strohmayer, T. E. 2005a, \apj,
329: in press (astro-ph/0509369).
330:
331: \bibitem[]{345} Bhattacharyya, S., \& Strohmayer, T. E. 2005b, in preparation.
332:
333: \bibitem[]{347} Bhattacharyya, S., Strohmayer, T. E., Miller, M. C., \&
334: Markwardt, C. B. 2005, \apj, 619, 483.
335:
336: \bibitem[]{350} Chakrabarty, D. et al. 2003, Nature, 424, 42.
337:
338: \bibitem[]{352} Cumming, A., \& Bildsten, L. 2001, \apj, 559, L127.
339:
340: \bibitem[]{354} Cumming, A., Morsink, S. M., Bildsten, L., Friedman, J. L.,
341: \& Holz, D. E. 2002, \apj, 564, 343.
342:
343: \bibitem[]{} Heyl, J. S. 2005, \mnras, 361, 504.
344:
345: \bibitem[]{} Lee, U., \& Strohmayer, T. E. 2005, \mnras, 361, 659.
346:
347: \bibitem[]{357} Miller, M. C. 1999, \apj, 515, L77.
348:
349: \bibitem[]{359} Miller, M. C. \& Lamb, F. K. 1998, \apj, 499, L37.
350:
351: \bibitem[]{361} Muno, M. P., Fox, D. W., Morgan, E. H., \& Bildsten,
352: L. 2000, \apj, 542, 1016.
353:
354: \bibitem[]{364} Muno, M. P., \"Ozel, F., \& Chakrabarty, D. 2002, \apj, 581,
355: 550.
356:
357: \bibitem[]{366} Nath, N. R., Strohmayer, T. E., \& Swank, J. H. 2002,
358: \apj, 564, 353.
359:
360: \bibitem[]{372} Press, W. H., Teukolsky, S. A., Vetterling, W. T., \&
361: Flannery, B. P. 1992, Numerical Recepies in FORTRAN (New York: Cambridge
362: University Press), 687-693.
363:
364: \bibitem[]{376} Spitkovsky, A., Levin, Y. \& Ushomirsky, G. 2002, \apj, 566,
365: 1018.
366:
367: \bibitem[]{378} Strohmayer, T.E., \& Bildsten, L. 2003, in {\it Compact
368: Stellar X-ray Sources}, Eds. W.H.G. Lewin and M. van der Klis,
369: (Cambridge University Press: Cambridge), (astro-ph/0301544).
370:
371: \bibitem[]{382} Strohmayer, T.E., Jahoda, K., Giles, A. B., \& Lee, U. 1997,
372: \apj,
373: 486, 355.
374:
375: \bibitem[11]{385} Strohmayer, T. E., \& Markwardt, C. B. 1999, \apj, 516, L81.
376:
377: \bibitem[]{387} Strohmayer, T. E., \& Markwardt, C. B. 2002, \apj, 577, 337.
378:
379: \bibitem[]{389} Strohmayer, T. E., Markwardt, C.B., Swank, J. H., \&
380: in 't Zand, J. 2003, \apj, 596, L67.
381:
382: \bibitem[]{392} Strohmayer, T. E., Zhang, W., \& Swank, J. H. 1997, \apj,
383: 487, L77.
384:
385: \bibitem[]{395} Watts, A. L., Strohmayer, T. E., \& Markwardt,
386: C. B. 2005, \apj, 634, 547.
387:
388: \end{thebibliography}{}
389:
390: \newpage
391:
392: \begin{deluxetable}{cccccccc}
393: \tablecolumns{8}
394: \tablewidth{0pc}
395: \tablecaption{Frequency evolution model parameters\tablenotemark{a} (with
396: 1$\sigma$ error)
397: for four bursts.}
398: \tablehead{ObsId & Start date & Burst & $\nu_{\rm 0}$ & $\dot\nu$ & $\ddot\nu$
399: & $\chi^{\rm
400: 2}$ & $Z^2_1$\tablenotemark{b}}
401: \startdata
402: 60032-05-03-00 & 2002 Jan 12 & A & $577.70^{+0.24}_{-0.25}$ &
403: $2.38^{+0.24}_{-0.25}$ &
404: $-0.42^{+0.07}_{-0.07}$ & $2.00$ & $170.42$ \\
405: \\
406: 60032-05-03-00 & 2002 Jan 13 & B & $579.19^{+0.07}_{-0.07}$ &
407: $0.81^{+0.06}_{-0.05}$ & --
408: & $7.16$ & $200.35$ \\
409: \\
410: 60032-05-05-00 & 2002 Jan 14 & C & $580.62^{+0.12}_{-0.12}$ &
411: $0.28^{+0.08}_{-0.08}$ & --
412: & $12.70$ & $173.18$ \\
413: \\
414: 60032-05-10-00 & 2002 Jan 22 & D & $578.44^{+0.12}_{-0.13}$ &
415: $1.21^{+0.11}_{-0.10}$ & -- &
416: $5.11$ & $89.75$ \\
417: \enddata
418: \tablenotetext{a}{Frequency evolution model: $\nu(t) = \nu_{\rm 0} + \dot\nu t
419: +
420: \ddot\nu t^2$, where $\nu_{\rm 0} = \nu(0).$}
421: \tablenotetext{b}{Fundamental power during burst rise.}
422: \end{deluxetable}
423:
424: \clearpage
425:
426: \begin{deluxetable}{ccccccc}
427: \tablecolumns{7}
428: \tablewidth{0pc}
429: \tablecaption{Fitting of phase-folded lightcurve\tablenotemark{a}: best fit
430: parameter\tablenotemark{b} values (with 1$\sigma$ error).}
431: \tablehead{Model & $a_{\rm 0}$ & $a_{\rm 1}$ & $\epsilon_{\rm 1}$ & $a_{\rm
432: 2}$ &
433: $\epsilon_{\rm 2}$ & $\chi^2$/dof}
434: \startdata
435: 1\tablenotemark{c} & $667.87\pm10.00$ & $241.42\pm14.23$ & $0.37\pm0.01$ & --
436: & -- &
437: $26.26/13$ \\
438: \\
439: 2\tablenotemark{d} & $668.99\pm10.00$ & $245.60\pm14.26$ & $0.37\pm0.01$ &
440: $59.49\pm14.05$ &
441: $0.01\pm0.02$ & $8.33/11$ \\
442: \enddata
443: \tablenotetext{a}{Combination (for first 1/3rd of the rise time interval)
444: of nine bursts from the following ObsIds (start date):
445: 30053-02-02-02 (1998 Aug 19), 30053-02-02-00 (1998 Aug 20), 40028-01-02-00
446: (1999 Feb 27),
447: 40028-01-08-00 (1999 Jun 18), 50030-02-01-00 (2000 Nov 05), 60032-05-03-00
448: (2002 Jan
449: 12), 60032-05-03-00 (2002 Jan 13), 60032-05-05-00 (2002 Jan 14) and
450: 60032-05-13-00
451: (2002 Feb 05).}
452: \tablenotetext{b}{Model lightcurve intensity is $I = a_{\rm 0} + a_{\rm 1}
453: \sin (2 \pi (\epsilon - \epsilon_{\rm 1})) + a_{\rm 2} \sin (4 \pi (\epsilon -
454: \epsilon_{\rm 2}))$, where $\epsilon$ is the phase variable.}
455: \tablenotetext{c}{Only fundamental.}
456: \tablenotetext{d}{Fundamental + 1st harmonic.}
457: \end{deluxetable}
458:
459: \clearpage
460: \begin{figure}
461: \epsscale{.6}
462: \plotone{f1.eps}
463: \caption{Time evolution of different observed burst properties during
464: the rise of burst A (see Table 1) from 4U 1636--536. Panel {\it a}
465: gives the detected intensity (histogram), power contours (minimum and
466: maximum power values are $18$ and $55$) using dynamic power spectra
467: (for 0.4 s duration at 0.02 s intervals), and the best fit model from
468: Table 1. Panel {\it b} gives the phase residuals (phase varies from 0
469: to 1) and rms deviation (broken horizontal lines) for constant
470: frequency $(580.72$~Hz) fitting. Panel {\it c} is same as panel {\it
471: b}, but for the best fit values of frequency evolution parameters
472: (Table 1). Panel {\it d} gives the rms amplitudes of brightness
473: oscillation (persistent emission subtracted). The upper histogram is
474: for the fundamental amplitude, and the lower histogram is for the 1st
475: harmonic amplitude. Here the horizontal lines give the binsize and the
476: vertical lines give $1\sigma$ error.}\end{figure}
477:
478: \clearpage
479: \begin{figure}
480: \epsscale{.7}
481: \plotone{f2.eps}
482: \caption{Similar as panel {\it a} of Fig. 1 (panel {\it a} is for
483: burst B, panel {\it b} is for burst C, and panel {\it c} is for burst
484: D). For each of the panels, the dynamic power spectra are calculated
485: for 0.3 s duration at 0.015 s intervals. The minimum and maximum power
486: values of the contours are $(23,77)$, $(20,53)$, and $(15,46)$ for
487: panels {\it a}, {\it b}, and {\it c} respectively.}
488: \end{figure}
489:
490: \clearpage
491: \begin{figure}
492: \epsscale{1.0}
493: \plotone{f3.eps}
494: \caption{Phase-folded lightcurve of burst oscillation during burst
495: rise from 4U 1636--536. The solid curve shows the data (combined for
496: nine bursts for 1/3rd of the rise time interval;
497: see Table 2), dashed curve is model 2 (see Table 2),
498: solid horizontal line is the constant level of the model, dotted curve
499: is the fundamental component of the model, and dash-dot curve is the
500: 1st harmonic component of the model.}
501: \end{figure}
502:
503: \end{document}
504: