astro-ph0404102/ms.tex
1: %%
2: %%  X-Ray Timing 2003 paper latex file.
3: %%
4: \documentclass[
5: %%   ,final            % use final for the camera ready runs
6:   ,draft            % use draft while you are working on the paper
7: %%  ,numberedheadings % uncomment this option for numbered sections
8: %%  ,                 % add further options here if necessary
9:   ]
10:   {aipproc}
11: 
12: \layoutstyle{8x11double}
13: 
14: % \bibliographystyle{aipproc}
15: %
16: %  In analogy with \leq and \geq, define relations for less (or greater)
17: %  than or approximately equal to:
18: %
19: \newcommand{\ltsim}{\lower.5ex\hbox{$\; \buildrel < \over \sim \;$}}
20: \newcommand{\gtsim}{\lower.5ex\hbox{$\; \buildrel > \over \sim \;$}}
21: \newcommand{\order}[1]{\mbox{$\cal{O}$ ({#1})}}
22: %
23: \newcommand{\etal}{\mbox{\it et~al.}}
24: \newcommand{\suph}{\mbox{$^{\rm h}$}}
25: \newcommand{\supm}{\mbox{$^{\rm m}$}}
26: \newcommand{\sups}{\mbox{$^{\rm s}$}}
27: 
28: \def\arcmin{\hbox{$^\prime$}}
29: \def\arcsec{\hbox{$^{\prime\prime}$}}
30: \def\degs       {$^\circ$}
31: \def\ergps      {$\rm erg$ ~\pers }
32: \def\fluxunit   {erg cm$^{-2}$s$^{-1}$keV$^{-1}$}
33: 
34: \newcommand{\units}[1]{\mbox{$\rm\,#1$}}
35: \newcommand{\kpc}       {\mbox{\rm\,kpc}}
36: \newcommand{\Hz}        {\mbox{\rm\,Hz}}
37: \newcommand{\kHz}       {\mbox{\rm\,kHz}}
38: \newcommand{\MHz}       {\mbox{\rm\,MHz}}
39: \newcommand{\GHz}       {\mbox{\rm\,GHz}}
40: \newcommand{\erg}       {\mbox{\rm\,erg}}
41: \newcommand{\Jy}        {\mbox{\rm\,Jy}}
42: \newcommand{\mJy}       {\mbox{\rm\,mJy}}
43: \newcommand{\uJy}       {\mbox{$\,\mu\rm{Jy}$}}
44: \newcommand{\K}         {\mbox{\rm\,K}}
45: \newcommand{\KperJy}    {\mbox{$\rm\,K\,Jy^{-1}$}}
46: \newcommand{\dm}        {\mbox{$\rm\,pc\,cm^{-3}$}}
47: \newcommand{\kms}       {\mbox{$\rm\,km\,s^{-1}$}}
48: \newcommand{\yr}        {\mbox{\rm\,y}}
49: \newcommand{\s}         {\mbox{\rm\,s}}
50: \newcommand{\ms}        {\mbox{\rm\,ms}}
51: \newcommand{\us}        {\mbox{$\,\mu\rm{s}$}}
52: \newcommand{\cm}        {\mbox{\rm\,cm}}
53: \newcommand{\m}         {\mbox{\rm\,m}}
54: \newcommand{\degyr}     {\mbox{$\,^\circ\,{\rm y}^{-1}$}}
55: \newcommand{\degree}    {\mbox{$^\circ$}}
56: \newcommand{\gauss}     {\mbox{\rm\,G}}
57: \newcommand{\accel}     {\mbox{$\rm\,m\,s^{-2}$}}
58: \newcommand{\Msun}      {\mbox{$\,M_{\mathord\odot}$}}
59: \newcommand{\Mearth}    {\mbox{$\,M_{\mathord\oplus}$}}
60: \newcommand{\Lsun}      {\mbox{$\,L_{\mathord\odot}$}}
61: \newcommand{\Rsun}      {\mbox{$\,R_{\mathord\odot}$}}
62: \newcommand{\OMC}		{\mbox{${\rm O}-{\rm C}$}}
63: \newcommand{\omczero}   {\mbox{${\rm O}-{\rm C}$}\,\,=\,\,0}
64: 
65: \newcommand{\Porb}{\mbox{$P_{\rm orb}$}}
66: \newcommand{\Ppul}{\mbox{$P_{\rm pul}$}}
67: \newcommand{\Jorb}{\mbox{$J_{\rm orb}$}}
68: \newcommand{\Porbdot}{\mbox{$\dot P_{\rm orb}$}}
69: \newcommand{\Ppuldot}{\mbox{$\dot P_{\rm pul}$}}
70: \newcommand{\Porbddot}{\mbox{$\ddot P_{\rm orb}$}}
71: \newcommand{\Ppulddot}{\mbox{$\ddot P_{\rm pul}$}}
72: \newcommand{\Rltwo}{\mbox{$R_{L2}$}}
73: \newcommand{\Rlone}{\mbox{$R_{L1}$}}
74: \newcommand{\Rltwodot}{\mbox{$\dot R_{L2}$}}
75: \newcommand{\Rlonedot}{\mbox{$\dot R_{L1}$}}
76: 
77: \newcommand{\exo}{\mbox{EXO\,0748$-$676}}
78: \newcommand{\xonesixfiveeight}{\mbox{X1658$-$298}}
79: \newcommand{\xoneighttwozero}{\mbox{X1820$-$303}}
80: \newcommand{\xoneighttwotwo}{\mbox{X1822$-$371}}
81: \newcommand{\xtwoonetwoseven}{\mbox{X2127$+$119}}
82: \newcommand{\smcxone}{\mbox{SMC\,X$-$1}}
83: \newcommand{\herxone}{\mbox{Her\,X$-$1}}
84: \newcommand{\grojoneseven}{\mbox{GRO\,J1744$-$28}}
85: \newcommand{\voph}{\mbox{V2301\,Oph}}
86: \newcommand{\amher}{\mbox{AM\,\,Her}}
87: 
88: \begin{document}
89: 
90: \title{X-Ray Eclipse Timing in the LMXB EXO0748-676}
91: 
92: \author{Michael T. Wolff}{address={E.O. Hulburt Center for Space Research, 
93: Naval Research Laboratory, Washington, DC 20735}}
94: \author{Paul S. Ray}{address={E.O. Hulburt Center for Space Research, 
95: Naval Research Laboratory, Washington, DC 20735}}
96: \author{Kent S. Wood}{address={E.O. Hulburt Center for Space Research, 
97: Naval Research Laboratory, Washington, DC 20735}}
98: 
99: \begin{abstract}
100: Orbital period changes are an important diagnostic for understanding 
101: low mass X-ray binary (LMXB) accretion-induced angular momentum 
102: exchange and overall system evolution.  
103: We present our most recent results for the eclipse timing of the 
104: LMXB \exo.  
105: Since its discovery in 1985 it has apparently undergone three distinct 
106: orbital period ``epochs", each characterized by a different 
107: orbital period than the previous epoch.  
108: We outline the orbital period behavior for \exo\ over 
109: the past 18 years and discuss the implications of this behavior 
110: in light of current theoretical ideas for LMXB evolution.
111: \end{abstract}
112: 
113: \maketitle
114: 
115: \section{Introduction}
116: 
117: There are eight currently known low mass X-ray binary (LMXB) systems
118: that undergo full or partial eclipses.
119: Such systems are important for the study of evolution in
120: LMXBs because the eclipse edges provide timing markers that make
121: possible the systematic observation of orbital period changes. 
122: The best studied orbit in this group is that of 
123: \exo\ which has been X-ray active since 1985 \citep{pwgg86}.
124: This long timeline has allowed an unprecedented look at its orbital 
125: dynamics \citep[see][and references therein]{whw+02}. 
126: The emerging picture for the orbital period behavior,
127: however, is anything but the expected smooth variations in \Porb\
128: and \Porbdot\ based on theoretical calculations 
129: done to date \citep[e.g., see][and references therein]{prp02}.
130: Rather, the observed period changes are discontinuous 
131: across multiple distinct epochs, 
132: and large apparent changes in \Porb\ of the order seconds 
133: can be observed on timescales as short as one orbit.
134: The magnitude of the observed changes in orbital period are 
135: much larger than expected from LMXB evolutionary theory.  
136: This is likely an indication that the observed variations in 
137: \Porb\ are short timescale effects of angular momentum 
138: redistribution in the system and are masking 
139: the underlying long term orbit evolution.
140: The Rossi X-Ray Timing Explorer (RXTE) satellite has allowed
141: us to monitor systematically the orbit of \exo\ in an effort to
142: delineate and understand the process of angular momentum exchange
143: and its effects on the \exo\ orbit. 
144: In this conference paper we report our recent progress. 
145: 
146: \section{RXTE Observations of \exo}
147: 
148: We now have 85 additional RXTE mid-eclipse time measurements 
149: over and above those analyzed in \citet{whw+02} for \exo. 
150: \citeauthor{whw+02} reported that the orbit period increased around 1990 
151: in a somewhat discontinuous manner by $\sim$7.9 ms since its discovery 
152: in February 1985. 
153: The cumulative data set of mid-eclipse timings considered in that 
154: paper ended in December 2000. 
155: Our present data set extends to observations done through October 5, 2003. 
156: \OMC\ (Observed - Calculated) diagrams for all published mid-eclipse 
157: timings of \exo\ plus the additional timings we 
158: present here (a total of 262 mid-eclipse time measurements) are shown 
159: in Figures~\ref{fig-exoallomc} and~\ref{fig-exorxteusaomc}. 
160: 
161: \begin{figure}
162: \includegraphics[height=4.5in,angle=270.0]{f1.ps}
163: \caption{
164: The \OMC\ residuals for all published mid-eclipse timings 
165: of \exo\ plus those additional timings we present here 
166: for a total of 262 mid-eclipse timings. 
167: All times are in Modified Julian Days corrected to the solar 
168: system barycenter. 
169: The measurement errors are as shown and except for a few cases 
170: are less than $\sim 1-2$ seconds. 
171: The solid line represents the Double-Broken Constant Period 
172: model for these data discussed in the text. 
173: The large deflection of the ASCA data away from either solution 
174: may be attributable to large excursions due to accumulated jitter 
175: as discussed in \citep{whw+02}. 
176: The systematic wandering of the orbit period caused by cumulative 
177: jitter prior to MJD 51710 is apparent. 
178: After this date, however, the orbit period takes on a new value 
179: that is only 1.4 ms larger than the apparent orbit period prior 
180: to MJD 48026. 
181: \label{fig-exoallomc}}
182: \end{figure}
183: 
184: \begin{figure}
185: \includegraphics[height=4.5in,angle=270.0]{f2.ps}
186: \caption{
187: The \OMC\ residuals for all USA and RXTE mid-eclipse 
188: timings of \exo\ plotted against the middle period of the 
189: Double-Broken Constant Period solution in Table~\ref{tbl-exoephem}.
190: All times are in Modified Julian Days corrected to the solar 
191: system barycenter. 
192: The solid line represents a constant period solution for the 
193: \OMC\ residuals  
194: and is clearly unacceptable as a model for these data. 
195: The systematic wandering of the orbit period caused by cumulative 
196: jitter prior to MJD 51710 is apparent. 
197: After this date, however, the orbit period takes on a new value 
198: of $\Porb = 0.15933774511(16) \times 10^{-10}$ (days)
199: and the wandering about the mean period is noticeably reduced.
200: \label{fig-exorxteusaomc}}
201: \end{figure}
202: 
203: The additional eclipse measurements display a remarkable behavior.  
204: When the eclipses observed after December 2000 are 
205: included in our analysis the orbit period appears to return to 
206: a value close to the orbit period observed during 
207: the EXOSAT/GINGA epoch.
208: We fit a three-constant-period model (\Porb$_0$, \Porb$_1$, 
209: and \Porb$_2$) to these data 
210: in which we constrain the phase to be constant across the 
211: instantaneous period changes but 
212: let the cycles ($n_{b0}$ and $n_{b1}$) of the 
213: period change be free parameters (see Table~\ref{tbl-exoephem}).
214: The best-fit model gives a positive 
215: period change of 
216: about \Porb$_1$ $-$ \Porb$_0$ $=$ $8.19 \pm 0.06$ ms similar 
217: to that reported in \citep{whw+02} 
218: during the GINGA to early RXTE era.
219: However, the additional data does not continue the trend of random 
220: wandering about the new period. 
221: Instead, the apparent \exo\ orbital period undergoes 
222: another change near MJD 51710 (near cycle 35136) 
223: of \Porb$_2$ $-$ \Porb$_1$ $=$ $- 6.75 \pm 0.02$ ms. 
224: This brings the net change in orbit period in 
225: the EXOSAT$-$RXTE eras (February 1985 through October 2003) to 
226: only \Porb$_2$ $-$ \Porb$_0$ $=$ $1.45 \pm 0.06$ ms, significantly 
227: less than reported in \citep{whw+02}.
228: Note that the orbit period in the present epoch (\Porb$_2$)
229: is distinct from the period in the EXOSAT/GINGA epoch (\Porb$_0$)
230: according to our analysis.
231: 
232: \section{Cumulative Jitter in X-Ray Mid-Eclipse Timings}
233: 
234: The \OMC\ residuals in Figure~\ref{fig-exorxteusaomc} appear 
235: correlated, that is, they 
236: are often systematically grouped above and below the zero residual line. 
237: In some cases the positive or negative variations in the mid-eclipse 
238: residuals can persist for many months. 
239: We have referred to this 
240: variation in the apparent orbital period as ``jitter", and apparent 
241: changes in orbital period can be caused by accumulated intrinsic 
242: jitter in a constant orbital period. 
243: ``Measurement error" is simply the random error associated with our 
244: measurement of the individual mid-eclipse times. 
245: We take ``intrinsic jitter" to mean the systematic wandering 
246: of the mid-eclipse timings around a smooth underlying model that 
247: can not be accounted for by random uncorrelated measurement errors. 
248: In particular, the systematic wandering of the mid-eclipse 
249: residuals apparent in Figure~\ref{fig-exorxteusaomc} can be 
250: accounted for by the accumulated intrinsic period jitter 
251: in individual orbit cycles. 
252: In order to investigate the observational measurement error 
253: inherent in our timing analysis and the intrinsic period jitter 
254: that is caused by other mechanisms (possibly mechanisms inside 
255: the LMXB system) we apply Maximum Likelihood Method \citep[MLM, see][]{k96} 
256: to estimate the parameters of a model for the orbital 
257: evolution that simultaneously includes an orbit period, an 
258: orbit period derivative, non-zero random intrinsic scatter 
259: (variance $\sigma_{\epsilon}^2$), and non-zero random 
260: measurement error (variance $\sigma_e^2$). 
261: We give results for \exo\ in two cases: (a) RXTE and USA data 
262: prior to MJD 51710, and, 
263: (b) the RXTE data after MJD 51710. 
264: A MLM model for the \exo\ RXTE and USA mid-eclipse timing data 
265: prior to MJD 51710 that includes a non-zero period derivative, 
266: intrinsic jitter and measurement error 
267: {\it is not} statistically preferred over a model with only 
268: intrinsic jitter and measurement error (a model with no intrinsic 
269: period jitter is strongly rejected according to the MLM analysis). 
270: The MLM analysis suggests that the intrinsic jitter
271: in the eclipse timings is characterized by $\sigma_{\epsilon} \sim 0.13$ s
272: with measurement error $\sigma_e \sim 1.3$ s,
273: similar to the results in \citep{whw+02}.
274: However, for the RXTE data after MJD 51710 the MLM analysis  
275: suggests that while the measurement error in the eclipse timings 
276: remains roughly constant at $\sigma_e \sim 1.2$ s, 
277: the intrinsic period jitter is greatly reduced 
278: to $\sigma_{\epsilon} \sim 0.028$ s.
279: Also, as was the case for the pre-MJD 51710 data, 
280: a model for the post-MJD 51710 timing data that 
281: includes a non-zero period derivative, intrinsic jitter 
282: and measurement error {\it is not} statistically preferred over a model 
283: with only intrinsic period jitter and measurement error.
284: Whatever process resulted in the abrupt change in the
285: orbital period after MJD 51710 also brought about a 
286: reduction in the magnitude of the intrinsic orbital period jitter
287: according to the MLM analysis.
288: 
289: \section{Discussion}
290: 
291: The X-ray eclipse timings for \exo\ appear to show three distinct 
292: orbital periods in three successive epochs 
293: along with significant intrinsic period jitter in 
294: these epochs. 
295: Furthermore, inspection of Figure~\ref{fig-exorxteusaomc} shows 
296: that during the early part of the RXTE epoch (until 
297: approximately March 2000) the period jitter is especially prominent. 
298: After this date the orbit period changes again by $\sim$-6.7 ms 
299: nearly returning to the value characteristic of
300: the EXOSAT$-$GINGA epoch. 
301: Such abruptly changing orbit periods are observed in a number 
302: of Algol binaries [see the discussion in \citep{simo99} and 
303: references found therein] although with 
304: larger amplitude in ${\Delta \Porb}/{\Porb}$. 
305: For Algol systems ${\Delta \Porb}/{\Porb} \sim 10^{-5}$ 
306: whereas in \exo\ we find ${\Delta \Porb}/{\Porb} \sim 10^{-6}$.
307: \citep{hall91} attributed the abrupt period changes in Algol systems 
308: to magnetic activity in a convective secondary star inducing 
309: changes in its quadrupole moment and altering the orbital 
310: angular momentum. 
311: If this explanation is applicable to the period changes 
312: in \exo\ \citep[see also][]{hwc97} then around the 
313: March 2000 magnetic activity in the secondary was altered in 
314: some manner changing the secondary's influence on the system orbital 
315: angular momentum distribution. 
316: Before this date the orbit period jittered around a mean value
317: but the jitter is reduced after this date.  
318: This may imply that magnetic activity in the secondary also was reduced, 
319: perhaps as a result of magnetic cycling similar to 
320: the 11-year cycle in the Sun. 
321: Small changes in radius of the secondary are predicted by the 
322: magnetic activity explanation of Algol orbit period changes put
323: forward by \citet{ap87} due to the changing magnetic pressure support 
324: for the outer layers of the secondary star. 
325: The data in Figure~\ref{fig-exoecldurvsomc}
326: \begin{figure}
327: \includegraphics[height=4.5in,angle=270.0]{f3.ps}
328: \caption{
329: The durations of all RXTE observed eclipses plotted 
330: as a function of observed \OMC\ residual for \exo. 
331: The points at increasing negative \OMC\ values 
332: are the same eclipses observed after the dramatic change
333: in period near MJD 51710 in Figure~\ref{fig-exorxteusaomc}.
334: \label{fig-exoecldurvsomc}}
335: \end{figure}
336: suggest that the radius of the
337: occulting star decreased across the MJD 51710 boundary. 
338: Totality duration is $\sim$7 seconds shorter after MJD 51710 
339: than before MJD 51710.
340: However, a spherically symmetric decrease of 1-2\% in
341: the radius of the secondary star is unlikely and other 
342: non-spherical changes in the radius
343: of the secondary must be considered as well.
344: 
345: The measurements shown in Figures~\ref{fig-exoallomc},~\ref{fig-exorxteusaomc}, 
346: and~\ref{fig-exoecldurvsomc}
347: give a magnitude and direction for the abrupt orbital period changes
348: and the magnitude and direction of the abrupt change in the
349: duration of X-ray totality.
350: Any attempt to understand these data will have to pose some
351: sort of almost instantaneous change in the binary
352: system parameters with an accompanying redistribution 
353: of the orbital and spin angular momenta rather than a slow 
354: change in the system because of the effects of mass exchange.
355: 
356: \begin{theacknowledgments}
357: This research is supported by the Office of Naval Research,
358: the NASA Astrophysical Data Program, and the 
359: NASA RXTE Guest Observer Program.
360: \end{theacknowledgments}
361: 
362: % \bibliography{journapj,usarefs}
363: \begin{thebibliography}{8}
364: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
365: \providecommand{\enquote}[1]{``#1''}
366: \expandafter\ifx\csname url\endcsname\relax
367:   \def\url#1{\texttt{#1}}\fi
368: \expandafter\ifx\csname urlprefix\endcsname\relax\def\urlprefix{URL }\fi
369: 
370: \bibitem[Parmar et~al.(1986)]{pwgg86}
371: Parmar, A.~N., White, N.~E., Giommi, P., and Gottwald, M., \emph{ApJ},
372:   \textbf{308}, 199 (1986).
373: 
374: \bibitem[Wolff et~al.(2002)]{whw+02}
375: Wolff, M.~T., Hertz, P.~L., Wood, K.~S., Ray, P.~S., and Bandyopadhyay, R.~M.,
376:   \emph{ApJ}, \textbf{575}, 384--396 (2002).
377: 
378: \bibitem[Podsiadlowski et~al.(2002)]{prp02}
379: Podsiadlowski, P., Rappaport, S., and Pfahl, E., \emph{ApJ}, \textbf{565}, 1107
380:   (2002).
381: 
382: \bibitem[Koen(1996)]{k96}
383: Koen, C., \emph{MNRAS}, \textbf{283}, 471 (1996).
384: 
385: \bibitem[Simon(1999)]{simo99}
386: Simon, V., \emph{A\&AS}, \textbf{134}, 1--19 (1999).
387: 
388: \bibitem[Hall(1991)]{hall91}
389: Hall, D.~S., \emph{ApJ}, \textbf{380}, L85--L87 (1991).
390: 
391: \bibitem[Hertz et~al.(1997)]{hwc97}
392: Hertz, P., Wood, K.~S., and Cominsky, L.~R., \emph{ApJ}, \textbf{486}, 1000
393:   (1997).
394: 
395: \bibitem[{Applegate} and {Patterson}(1987)]{ap87}
396: {Applegate}, J.~H., and {Patterson}, J., \emph{ApJ}, \textbf{322}, L99--L102
397:   (1987).
398: 
399: \end{thebibliography}
400: 
401: 
402: \begin{table}
403: \begin{tabular}{lcc}
404: \hline
405: \tablehead{1}{l}{b}{Parameter}
406: &\tablehead{1}{c}{b}{}
407: &\tablehead{1}{c}{b}{Value}\\
408: \hline
409: \multicolumn{3}{@{}c}{\bfseries All Eclipse Timing Data}\\
410: \hline
411: $T_0$ (MJD/TDB)		&$=$	&$46111.0752010(42)$\\
412: $\Porb_{0}$ (day)	&$=$	&$0.15933772838(66)$\\
413: $n_{b0}$ (cycle)	&$=$	&$12019.5(63.8)$\\
414: $\Porb_{1}$ (day)	&$=$	&$0.15933782322(10)$\\
415: $n_{b1}$ (cycle)	&$=$	&$35136.4(11.4)$\\
416: $\Porb_{2}$ (day)	&$=$	&$0.15933774511(16)$\\
417: $\chi^{2}$(dof)		&$=$	&$46.1 ( 256 )$\\
418: \hline
419: \end{tabular}
420: \caption{Orbital Ephemerides of \exo}
421: \label{tbl-exoephem}
422: \end{table}
423: 
424: \end{document}
425: