1: \documentstyle[epsfig,aas2pp4]{article}
2:
3: \begin{document}
4:
5: \title{Changes in the long term intensity variations in Cyg~X-2~and~LMC~X-3}
6:
7: \author{B. Paul$^{1,}\footnotemark[3]$, S. Kitamoto$^2$ and F. Makino$^1$ }
8: \affil{$^1$Institute of Space and Astronautical Science,
9: 3-1-1 Yoshinodai, Sagamihara, Kanagawa 229-8510, Japan}
10: \affil{$^2$Department of Earth and Space Science, Faculty of Science,
11: Osaka University 1-1, Machikaneyama, Toyonaka, Osaka, 560, Japan}
12: \affil{e-mail:~~bpaul@astro.isas.ac.jp, kitamoto@ess.sci.osaka-u.ac.jp,
13: makino@astro.isas.ac.jp}
14: \footnotetext[3]{On leave from the Tata Institute of Fundamental
15: Research, Homi Bhaba Road, Mumbai, 400005, India}
16:
17: \begin{abstract}
18:
19: We report the detection of changes in the long term intensity
20: variations in two X-ray binaries Cyg X-2 and LMC X-3. In this work, we have
21: used the long term light curves obtained with the All Sky Monitors (ASM) of
22: the Rossi {\it X-ray Timing Explorer (RXTE), GINGA, ARIEL 5} and {\it VELA 5B}
23: and scanning modulation collimator of {\it HEAO~1}. It is found that in the
24: light curves of both the sources, obtained with these instruments at various
25: times over the last 30 years, more than one periodic or quasi-periodic
26: components are always present. The multiple prominent peaks in the
27: periodograms have frequencies unrelated to each other. In Cyg X-2, {\it
28: RXTE-ASM} data show
29: strong peaks at 40.4 and 68.8 days, {\it GINGA-ASM} data show strong peaks at
30: 53.7 and 61.3 days. Multiple peaks are also observed in LMC X-3. The various
31: strong peaks in the periodograms of LMC X-3 appear at 104, 169 and 216
32: days with {\it RXTE-ASM}, and 105, 214 and 328 days with {\it GINGA-ASM}. The
33: present results, when compared with the earlier observations of periodicities
34: in these two systems, demonstrate the absence of any stable long period. The
35: 78 day periodicity detected earlier in Cyg X-2 was probably due to the
36: short time base in the {\it RXTE} data that were used and the periodicity of
37: 198 days in LMC X-3 was due to a relatively short duration of observation with
38: {\it HEAO~1}.
39:
40: \end{abstract}
41:
42: \keywords{accretion, accretion disks -- stars: individual (Cyg X-2, LMC X-3)
43: -- stars: neutron -- X-rays: binaries}
44:
45: \section{Introduction}
46:
47: Many X-ray binaries are highly variable
48: in their X-ray intensity over long time scales. But in most of the sources,
49: either the intensity variations are aperiodic or their periodic nature has not
50: yet been discovered because of long-period, low modulation or a lack of
51: sensitive uninterrupted monitoring. The All Sky Monitor (ASM) onboard the Rossi
52: {\it X-ray Timing Explorer (RXTE)} has produced light curves of many bright
53: X-ray binaries for about 1,000 days with large signal to noise ratio. The
54: {\it RXTE-ASM} data along with the {\it GINGA-ASM, VELA 5B, ARIEL 5} and
55: {\it HEAO~1} data have a large time base and it is possible not only to
56: search for any periodic or quasi periodic intensity variations of a few days
57: to a few months time scale, but also to investigate changes in
58: the timing behaviour. Changes in the long term periodicity have been observed
59: in SMC X-1 (Wojdowski et al. 1998) and GX 354-0 (Kong et al. 1998). We
60: selected Cyg X-2 and LMC X-3 to study any possible changes in the long
61: term periodicity because in these two sources long term periodicity is
62: known to exist in data bases covering a very long time and also for the
63: fact that light curves for the intervening period 1987--1991 were available
64: from the {\it GINGA-ASM}.
65:
66: Cyg X-2, a bright X-ray source, was discovered by Byram et al. (1966)
67: with a sounding rocket experiment. A binary period of 9.8 days, other
68: orbital parameters and mass limits of the two components of this binary
69: system were measured by Cowley et al. (1979). This is a low mass X-ray
70: binary and a typical Z source that shows Type-I X-ray bursts (Kahn \&
71: Grindlay 1984; Smale 1998), 18$-$50 Hz QPOs in the horizontal branch
72: and 5.6 Hz QPOs in the nominal branch (Hasinger et al. 1986; Elsner et
73: al. 1986; Wijnands et al. 1997; Kuulkers, Wijnands and van der Klis 1999).
74: Recently, with the {\it RXTE}, kHz
75: QPOs with two peaks have also been detected from this source (Wijnands
76: et al. 1998). Vrtilek et al. (1988) discovered X-ray dips of different
77: types and some were found to occur in one particular phase of the binary
78: period. A 77 day periodicity was discovered from the {\it VELA 5B }
79: observations (Smale \& Lochner 1992). Wijnands, Kuulkers \& Smale (1996)
80: reported detection of a periodicity of 78 days from the first 160 days
81: of ASM data, and showed that their result was supported by the archival
82: data of {\it ARIEL 5} and {\it VELA 5B}.
83:
84: LMC X-3, a high mass X-ray binary, was discovered with the {\it UHURU}
85: satellite (Leong et al. 1971) and its position was measured accurately with
86: the {\it HEAO~1} scanning modulation collimator (Johnston et al. 1978). From
87: spectroscopic observations Cowley et al. (1983) discovered an orbital
88: period of 1.7 days and a mass function of 2.3 M$_\odot$. LMC X-3 is
89: considered to be a very strong black hole candidate (BHC) due to the fact that
90: the mass of the compact object has a lower limit of 9 M$_\odot$. But, {\it
91: HEAO~A2} observations revealed that unlike other known black hole candidates,
92: LMC X-3 lacks rapid X-ray variability (Weisskopf et al. 1983). The lowest
93: time scale for 1\% rms amplitude variation was derived from the {\it EXOSAT}
94: observations to be 600~s (Treves et al. 1988). One possible
95: explanation for the lack of rapid X-ray variability is that most of the
96: time LMC X-3 is found to be in a high state with an unusually soft X-ray
97: spectrum (White \& Marshal 1984), and the rapid X-ray variations are
98: generally subdued in Cyg X-1 like BHCs in their high state. {\it GINGA}
99: observations showed that the energy spectrum consists of a soft, thermal
100: component and a hard, power-law component. In spite of large changes in
101: mass accretion rate and disk temperature, the inner radius of the
102: accretion disk was found to be remarkably constant and was suggested to
103: be related to the mass of the compact object (Ebisawa et al. 1993).
104: Long term intensity variation is well known in LMC X-3 and Cowley et
105: al. (1991) discovered a periodicity of 198 (or 99) days from
106: the {\it HEAO~A1} and {\it GINGA} large area counter data.
107:
108: The light curves of Cyg X-2 and LMC X-3 obtained with the ASMs onboard the
109: {\it RXTE, GINGA, VELA 5B} (only for Cyg X-2), {\it ARIEL 5} and the scanning
110: modulation collimator of the {\it HEAO~1} (only for LMC X-3) have been used
111: to investigate the long term intensity variations in these two sources and
112: we have found presence of multiple components of flux variations in both the
113: sources. The data used here together covers a time base of 30 years for
114: Cyg X-2 and 26 years for LMC X-3.
115:
116: \section{Data and analysis}
117:
118: The ASM on board {\it RXTE} scans the sky in a series of dwells of about 90 s
119: each,
120: and any given X-ray source is observed in about 5$-$10 such dwells every day.
121: The details about the ASM detectors and observations with the ASM are
122: given in Levine et al. (1996). We have used the quick look ASM data
123: obtained in the period 1996 February 20 to 1999 February 4, provided by the
124: {\it RXTE-ASM} team. The ASM data are available in two different forms, per
125: dwell and one day average. The periodograms obtained from the
126: two sets of data are identical except for a normalization. Hence the
127: results obtained from the one day average data are presented here. The {\it
128: GINGA-ASM} was operational during 1987 to 1991 and details of the detectors,
129: operations, and the detection techniques have been described by Tsunemi et
130: al. (1989). The all sky observations were performed about once every day when
131: the satellite was given one full rotation with about 70\% sky coverage by
132: the ASM. The detection limit in one such rotation was of the order of 50
133: mCrab and was dependent on the position of the source with respect to the
134: satellite equator. Archival data of {\it VELA 5B} and {\it ARIEL 5} were
135: obtained from the HEASARC data base and {\it HEAO~1} data for LMC X-3 were
136: taken from previously published work (Cowley et al. 1991). For details about
137: the {\it ARIEL 5} ASM and the {\it VELA 5} instrument please refer to
138: Holt (1976) and Priedhorsky, Terrel \& Holt (1983) respectively.
139:
140: To search for periodicities in the unevenly sampled data, we have used the
141: method suggested by Lomb (1976) \& Scargle (1982). For Cyg X-2, we generated
142: periodograms in the period range of 10$-$160 days from the {\it RXTE-ASM,
143: GINGA-ASM, ARIEL 5} and {\it VELA 5B} data. The {\it ARIEL 5, VELA 5B} and
144: initial part of the {\it RXTE-ASM} light curves were analyzed earlier and
145: long periods of 78 and 69 days were reported (Smale \& Lochner 1992;
146: Wijnands et al. 1996, Kong et al. 1998). For LMC X-3, light curves from
147: {\it RXTE-ASM, GINGA-ASM, ARIEL 5} and {\it HEAO~1} were used to generate
148: periodogram in the range of 20$-$500 days. {\it VELA 5B} data for LMC X-3
149: has low signal to noise ratio and is not used here. The discovery of a 198
150: (or 99) day periodicity from the {\it HEAO 1} light curve was made by Cowley
151: et al. (1991), and the initial part of the {\it RXTE-ASM} data also showed
152: similar variation (Wilms et al. 1998). The period ranges chosen for the
153: periodogram analysis are such that a sub-harmonic or first harmonic of the
154: earlier known periods can be identified for the respective sources. For
155: Cyg X-2, we also verified that in the 160-500 days range the periodograms
156: generated from data sets are featureless except for {\it ARIEL 5} which
157: shows peaks around 180 and 360 days due to known yearly effect (Priedhorsky
158: et al. 1983).
159:
160: \begin{figure}[t]
161: \centerline{\psfig{figure=fig1.ps,width=2.2in,angle=-90}}
162: \caption{The {\it RXTE-ASM} (1.5--12 keV) and {\it GINGA-ASM}
163: (1--20 keV) light curves of Cyg X-2. The days of the observations are given
164: in truncated Julian days in the figure.}
165: \end{figure}
166:
167: The significance of the various peaks detected in the periodograms is given
168: in Table 1 in terms of the false alarm probability (FAP) calculated following
169: the
170: method suggested by Horne \& Baliunas (1986). The window functions, which
171: can induce artificial periodicities were calculated for the different time
172: series used here. The {\it GINGA-ASM} time series, which is very susceptible
173: to artificial periodicities because of its scarce sampling, does not show
174: any alarming feature in the period
175: range used in this paper and the effect of the window function is less for
176: the other instruments. However, we wish to point out that the use of
177: window function to find spectral leakage can sometimes be misleading for this
178: kind of light curves. As an example, in the {\it HEAO 1} light curve of LMC X-3,
179: the density of data points available is larger when the source is brighter,
180: and this is likely to be true for other sources and instruments also. This
181: results in some features in the window function at periods near 80 and 160 days,
182: whereas a glance at the light curve (Figure 4) leaves no doubt about the
183: presence of a periodic variation. We have verified the significance of the
184: peaks in the periodograms independently using two more methods. Following
185: Kong et al. (1998), we have generated light curves using random numbers, with
186: the time series, average, and variance similar to the real light curves and
187: calculated the periodograms for 10,000 such light curves. The highest points
188: in these periodograms were identified and a power which is larger than the
189: highest points of 99\% of the periodograms is considered to correspond to a
190: false alarm probability of 10$^{-2}$. The same process was carried out also
191: for 10,000 light curves with the same time series as the real light curves,
192: but the count rates redistributed randomly. The results from these two
193: analysis are found to be identical and the 10$^{-2}$ false alarm levels are
194: indicated by dashed lines in all the periodograms. Absence
195: of strong peaks in the periodograms calculated from the simulated and
196: redistributed light curves confirms that the peaks observed in the periodograms
197: of the real light curves are not artifact of observation windows.
198:
199: \subsection{Cyg X-2}
200:
201: Presence of strong intensity variations by a factor of $\sim$2--4 on time
202: scale of weeks is well known in this source (Kuulkers, van der Klis, \&
203: Vaughan 1996; Wijnands et al. 1996) and can be clearly seen in the light
204: curves of {\it RXTE-ASM} and {\it GINGA-ASM} (Figure 1). A periodic nature of
205: this intensity variation with period of less than 100 days is also apparent
206: in the light curves.
207:
208: The periodograms obtained from the {\it RXTE-ASM, GINGA-ASM, ARIEL 5} and
209: {\it VELA 5B} data are shown in Figure 2. In the RXTE-ASM data, there are
210: several significant peaks and the two most prominent ones are at 40.4 and
211: 68.8 days indicating the presence of multiple periodicities in this system.
212: In the {\it GINGA-ASM} periodogram there are two prominent peaks at
213: 53.7 and 61.3 days and there are several smaller peaks. The periodogram of
214: the {\it ARIEL 5} data is a very complicated one and has many prominent
215: peaks at frequencies ranging between 30 and 80 days. In addition to
216: the 78 days peak noticed by Wijnands et al. (1996), there are other peaks at
217: 41, 46 and 54 days. The {\it VELA 5B} data has low signal to noise
218: ratio and the power in the periodogram is much smaller, but two peaks at 33.9
219: and 77.4 days can be clearly seen in Figure 2. The significance of the peaks
220: detected in various periodograms are given in Table 1.
221:
222: \begin{figure}[t]
223: \centerline{\psfig{figure=fig2.ps,width=4.2in,angle=-90}}
224: \caption[fig2.ps]{The Lomb-Scargle periodograms obtained from the
225: {\it RXTE-ASM, GINGA-ASM, ARIEL 5} and {\it VELA 5B} light curves of
226: Cyg X-2. The horizontal dashed line represent the 99\% confidence limits
227: (see text for details).}
228: \end{figure}
229:
230: To investigate whether these multiple peaks in the periodograms originate
231: in different parts of the light curves or whether multiple peaks are present
232: throughout the entire light curve of each satellite, we have divided the
233: light curves
234: of {\it RXTE} and {\it ARIEL 5} into three equal segments and generated
235: periodograms from each of them. There are several prominent peaks in each of
236: the periodograms, and the periods of these peaks are noted in Table 1 along
237: with the significance of their detection. Periodograms obtained from the three
238: 358 day segments of the {\it RXTE} light curve are shown in Figure 3
239: along with the respective light curves. It is evident from Figure 3 and Table
240: 1 that there are multiple periodicities in this system that are uncorrelated
241: and have varying amplitude and period. The higher frequency variation at
242: around 40 days time scale appears to be more stable in frequency than the
243: lower frequency variations around 70 days during these observations.
244:
245: \begin{figure}[t]
246: \centerline{\psfig{figure=fig3.ps,width=2.4in,angle=-90}}
247: \caption[fig3.ps]{Three segments of the {\it RXTE-ASM} light curve of
248: length 358 days each and the corresponding Lomb-Scargle periodograms are
249: shown on the left and right hand sides respectively. The 99\% confidence
250: limits are shown with the dashed horizontal lines.}
251: \end{figure}
252:
253: \subsection{LMC X-3}
254:
255: Long term light curves of LMC X-3, obtained with the {\it RXTE-ASM, GINGA-ASM}
256: and {HEAO~1} are shown in Figure 4. The {HEAO~1} data, plotted in the lower
257: panel of the figure, very clearly shows almost periodic intensity variations
258: at about 100 days. However, as Cowley et al. (1991) pointed out, the intensity
259: modulation is missing during the first 100 days. They concluded from this data
260: that the intensity variations in LMC X-3 is periodic at $\sim$ 198 (or
261: possibly $\sim$ 99) days. The {\it GINGA-ASM} light curve shown in the middle
262: panel of the figure also indicates intensity modulations, but at a larger time
263: scale of about 200 days. Two such modulations can be clearly identified at the
264: end of the light curve and one more in the middle. There are also some episodes
265: of about 100 days periodic variations in some parts of the {\it GINGA-ASM} light
266: curve. The {\it RXTE-ASM} light curve of LMC X-3, as shown in the top panel
267: of Figure 4, shows a varying nature of the approximately 100 or 200 days
268: intensity variations. At the beginning, there are three strong modulations at
269: about 100 days, similar to what was seen with {\it HEAO~1}, but in the later
270: part of the light curve, the modulations are much less prominent and have
271: larger time scale. Intensity variations in LMC X-3 are not clearly visible in
272: the {\it ARIEL 5} light curve (not shown here) because of sparse sampling.
273:
274: \begin{figure}[t]
275: \centerline{\psfig{figure=fig4.ps,width=2.2in,angle=-90}}
276: \caption[fig4.ps]{The long term light curves of LMC X-3 obtained with
277: the {\it RXTE-ASM, GINGA-ASM} and {\it HEAO~1} satellites. The energy ranges
278: are 1.5--12 keV, 1--20 keV and 1--13 keV respectively.}
279: \end{figure}
280:
281: In Figure 5 we have shown the periodograms generated from the light curves of
282: LMC X-3 obtained with all the four satellites mentioned above. There are three
283: prominent peaks at 104.4, 168.8 and 215.6 days in the periodogram obtained
284: from the {\it RXTE-ASM} light curve. The peak at 104.4 days is narrow while
285: the other two are broad. In the periodogram of the {\it GINGA-ASM}, however,
286: there are two prominent peaks at 214 and 328 days. There is also a less
287: significant indication of some periodic component at 105 days. The {\it HEAO~1}
288: data which have been extensively discussed by Cowley et al. (1991), predictably
289: shows two peaks at 99 and 203 days with the former being narrower. Results
290: from the {\it RXTE} and {\it HEAO~1} are similar in nature except for the fact
291: that the broad peak near 200 days is resolved into two components in the {\it
292: RXTE-ASM} data. This also shows that the two peaks near 100 and 200 days as
293: seen in the periodogram of the {\it HEAO~1} data are not related. The
294: periodogram of the {\it GINGA-ASM} light curve also has a shape similar to the
295: other two but the time scale is a factor of two larger. The {\it ARIEL 5}
296: light curve has low signal to noise ratio and infrequent sampling and the
297: periodogram obtained from this data shows three less significant peaks at 90,
298: 96 and 130 days. At periods below 50 days, this periodogram is very noisy.
299: From the four periodograms shown in Figure 5 it appears that there is a quasi
300: periodic component at around 100 days in LMC X-3 whose strength is time
301: dependent. There is at least one more component at longer period of about
302: 200 days. The period excursion of the $\sim$100 days component is significant,
303: but relatively smaller than the other periodicities.
304:
305: \begin{figure}[t]
306: \centerline{\psfig{figure=fig5.ps,width=4.2in,angle=-90}}
307: \caption[fig5.ps]{The Lomb-Scargle periodograms generated from the light
308: curves of LMC X-3 obtained with the {\it RXTE-ASM, GINGA-ASM, HEAO~1} and
309: {\it ARIEL 5} detectors. The dashed horizontal lines indicate the 99\%
310: confidence limits.}
311: \end{figure}
312:
313: To investigate the nature of the multiple components of intensity variations
314: in more detail, we have done further analysis of the {\it RXTE-ASM} data of
315: LMC X-3 in a manner similar to what was done with the Cyg X-2 data. We divided
316: the light curve into two segments, each 540 days long, and have generated the
317: periodograms from both of the segments that are shown in Figure 6.
318: Two large peaks at 104.3 and 177.8 days are clearly seen in the first
319: periodogram whereas the periodogram generated from the second part of the
320: light curve shows two peaks at 101.9 and 221.6 days with much less strength.
321: It appears that the $\sim$104 day periodicity is suppressed in the second
322: part of the {\it RXTE} data and the second component of intensity variation
323: has moved to a longer time scale. The significance of the peaks in the
324: periodograms are given in Table 1.
325:
326: \begin{figure}[t]
327: \centerline{\psfig{figure=fig6.ps,width=2.0in,angle=-90}}
328: \caption[fig6.ps]{The Lomb-Scargle periodograms obtained from two segments
329: of the {\it RXTE-ASM} light curve of LMC X-3. Each segment of the light curve
330: is of 540 days duration. The dashed horizontal lines indicate the 99\%
331: confidence limits.}
332: \end{figure}
333:
334: \subsection{Spectral variations}
335:
336: In Cyg X-2, the hardness ratio and intensity that define its position in
337: the Z track and also the position of the Z track itself, changes significantly
338: at short time scales. But at longer time scales, in between different
339: observations, a negative correlation was found between the hardness ratio and
340: total intensity with {\it EXOSAT} (Kuulkers et al. 1996), {\it GINGA}
341: (Wijnands et al. 1997) and also with the {\it RXTE-ASM} during its first few
342: months observations (Wijnands et al. 1996). The hardness ratio in LMC X-3, on
343: the other hand was found to have positive correlation with intensity (Cowley
344: et al. 1991). We have calculated the two hardness ratios HR1 (3.0-5.0
345: keV/1.5-3.0 keV) and HR2 (5.0-12 keV/3.0-5.0 keV) from the {\it RXTE-ASM}
346: data as a function of the total intensity for the two sources (Figure 7)
347: using all the available data. In Cyg X-2, HR2 is negatively correlated
348: with luminosity (correlation coefficient -0.6 and probability of no correlation
349: 10$^{-30}$) but HR1 does not show any correlation (coefficient -0.04,
350: probability 0.5). If the data points are connected by lines, the HR2 plot
351: appears like a loop indicating that HR2 follows two different tracks during
352: the rising and decaying phases of the intensity variations. Large deviation
353: in the hardness plot of Cyg X-2 is due to movement of the source along the Z
354: track (Wijnands et al. 1996) and changing position of the Z track in the
355: color-color diagram. In LMC X-3, a weak positive correlation is
356: found for HR1 (0.5, 10$^{-5}$) but no correlation for HR2 (-0.07, 0.5). The
357: correlation coefficients and probabilities were calculated using two different
358: methods: the linear and rank correlation, both gave identical results.
359:
360: \begin{figure}[t]
361: \centerline{\psfig{figure=fig7.ps,width=3.2in,angle=00}}
362: \caption[fig7.ps]{Hardness ratios HR1 (3-5 keV / 1.5-5 keV) and HR2
363: (5-12 keV/ 3-5 keV) of Cyg X-2 and LMC X-3 obtained from the {\it RXTE-ASM}
364: data are plotted against the total intensity. Cyg X-2 data points are averaged
365: for 3 days and LMC X-3 data points are averaged for 10 days.}
366: \end{figure}
367:
368:
369: \section{Discussion}
370:
371: Apart from the binary period, long term periodic variations are known to be
372: present in many X-ray binaries. There are four sources in which the presence
373: of long periods is very well established, (1) Her X-1, a 1.7 day binary with a
374: 35 day period (Giacconi et al. 1973), (2) LMC X-4, a 1.4 day binary with a
375: 30.5 day period (Lang et al. 1981), (3) SMC X-1, a 3.9 day binary with a
376: long period of 60 days (Wojdowski et al. 1998) and (4) SS 433, a 13.1 day
377: binary with a long period of 164 days (Margon et al. 1979).
378: Incidentally, the first three of these sources are also X-ray pulsars. In
379: SS 433, the periodicities are detected photometrically and spectroscopically
380: in the optical band only. There is also evidence of periodic component in
381: several other sources. Among the high mass X-ray binaries, periodicity has been
382: observed in Cyg X-1 at 300 days (Priedhorsky et al. 1983, also see Kitamoto
383: et al. (1999) for the {\it GINGA-ASM} observations of a $\sim$150 day period),
384: 4U 1907+09 at 42 days (Priedhorsky \& Terrell 1984)
385: and LMC X-3 at 198 (or 99) days (Cowley et al. 1991). Among the low
386: mass X-ray binaries, Smale \& Lochner (1992) found periodicity in three
387: sources, Cyg X-2 (78 days), 4U 1820-303 (175 days) and 4U 1916-053 (199 days).
388: A 106 day periodicity was discovered from an extragalactic point source in
389: the spiral galaxy M33 (Dubus et al. 1997). The {\it RXTE-ASM} observations of
390: a large number of sources, for the past three years, detected intensity
391: variations in many sources (see Levine 1998 for a summary and the light
392: curves). Some of these sources are of periodic nature and there has also been
393: discovery of new periodic sources using the {\it RXTE-ASM} data (Sco X-1,
394: Peele \& White 1996; X2127+119, Corbet, Peele \& Smith 1996). Unstable
395: long-term periodicity that is attributed to activity of the companion star
396: or instability of the accretion disk has been observed in Aql X-1
397: (Kitamoto et al. 1993). The ratio of the long and orbital periods in
398: these sources has a wide range between 5 (in 4U 1907+09) and 22,000 in (4U
399: 1820-303). In the two binaries Cyg X-2 and LMC X-3, the reported long term
400: periodicity is larger by a factor of 8 and 116 than the respective orbital
401: periods.
402:
403: The long periods in Her X-1, LMC X-4 and SS 433 are believed to be produced
404: by the precession of the accretion disks. The mechanisms proposed to cause the
405: precession of the disks are, (1) forced precession of a tilted disk by the
406: gravitational field of the companion star (Katz 1973) and (2) precession of a
407: disk that is slaved to a misaligned companion star (Roberts 1974). But, the
408: time scale of precession expected in the sources in which binary parameters
409: are well known is in disagreement with the observed long periods (Priedhorsky
410: \& Holt 1987). One additional problem is that a large excursion is observed
411: in the long period of Her X-1 (\"Ogelman 1985) and SS 433 (Margon 1984), and it
412: may also be present in LMC X-4. This is not explained in the above two models
413: of disk precession (see Priedhorsky \& Holt 1987, for a detailed discussion).
414: However, in a realistic case, when various factors like magnetic pressure,
415: radiation pressure, tidal force, relativistic frame dragging etc. are
416: considered, it is possible to have significant deviation in the precession
417: period from its time averaged value. Recent developments in the models,
418: including tilted
419: and twisted disks due to coronal winds (Schandl \& Meyer 1994; Schandl 1996)
420: and warped precessing disks due to radiation pressure (Maloney, Begelman \&
421: Pringle 1996; Wijers \& Pringle 1998) have a provision for variations in
422: the precession period. Other models have also been considered to explain
423: the long periods. These are 1) precession of the compact object (Tr\"umper et
424: al. 1986), 2) influence of a third body (Fabian et al. 1986) and 3) periodic
425: modulation of the mass accretion rate (Priedhorsky \& Holt 1987). Among
426: these models, the first one is not applicable for black hole sources,
427: and the second one appears to be not true for the two pulsars Her X-1
428: (Tananbaum et al. 1972) and LMC X-4 (Pietsch et al. 1985). In Her X-1 and LMC
429: X-4, the presence of a third body required to produce the long period, would
430: have resulted in additional detectable variations in the pulsation property
431: (Priedhorsky \& Holt 1987). Periodic and asymmetric mass transfer, which is
432: induced by the disk's shadowing of the Roche-lobe overflow region, also
433: contributes to the slaved nature of the accretion disk and is a possible
434: mechanism. Mass transfer feedback induced by X-ray irradiation may also
435: generate the observed long term periodicities (Osaki 1985).
436:
437: No clear understanding of the reason behind the observed
438: long-term periodicities stands out among all these possibilities.
439: None of the possibilities mentioned above can explain the long-term
440: periodicity in all the sources. In the high mass X-ray binaries, the
441: periodicity is
442: generally believed to be related to disk precession and in the low mass
443: X-ray binaries another possibility is some type of disk instability or
444: modulation in the mass accretion rate related to or induced by the
445: X-ray radiation (Meyer 1986).
446:
447: Varying obscuration by the disk (caused by precession), which provides a good
448: explanation for the long term periodicity, should have the following
449: observational consequence. With increased absorption in the low intensity
450: phase, a hardening in the spectrum is expected which is known to be present
451: in Her X-1 and LMC X-4. The spectral hardening is likely to be more pronounced
452: between the two lower energy bands. Contrary to this expectation, we find that
453: in Cyg X-2, HR1 is uncorrelated to the total intensity whereas HR2 is
454: anti-correlated, and in LMC X-3, HR1 has a weak positive correlation (Figure
455: 7). Therefore, the spectral variations do not provide very good support to
456: the disk obscuration scenario.
457:
458: The present work involving light curves of these two sources with very
459: large time base suggests that there is no stable periodicity in either of
460: these systems. There are indications that the oscillatory components seen
461: in the light curves have varying amplitude and period. Forced or slaved
462: precession of the accretion disk is unlikely to be the mechanism behind
463: the quasi periodic intensity variations observed here. Scenarios including
464: disk precession induced by radiation pressure or tilted and twisted disk
465: structure induced by wind also require spectral variations different from
466: the observed pattern. The presence of a third body is also likely to produce
467: more regular patterns in the intensity variations. Another plausible
468: explanation for the observed behaviour is instability in the disk or in the
469: mass accretion rate. But Kuulkers et al. (1996) have pointed out that changes
470: in the mass accretion rate is unlikely to produce the long term behaviour
471: observed in Cyg X-2, because changes in the mass accretion rate on a time
472: scale of less than a day and associated spectral changes are in fact known
473: to produce the Z pattern (Hasinger \& van der Klis 1989).
474:
475: However,
476: if we look only at the {\it RXTE} ASM data of Cyg X-2 (Figure 3), it appears
477: that there are two unrelated components of intensity variations at periods of
478: $\sim$40 and $>$60 days, with the first component more stable in period than
479: the second one. The results obtained from LMC X-3 (Figures 5 and 6) is somewhat
480: similar with a relatively stable component at $\sim$100 days and a highly
481: varying component at $>$130 days. It is possible that in both objects
482: there are two sources of intensity variation, the first one relatively
483: stable with a smaller period caused by precession of the accretion disk or by
484: a third body and the second one unstable in time and having a longer time
485: scale caused by disk instability or changes in the mass accretion rate.
486: Coupling between the two components may result in incorrect measurement
487: of the period of the first component.
488:
489: \begin{acknowledgements}
490:
491: We thank an anonymous referee for many suggestions which helped to improve
492: a previous version of the manuscript.
493: This research has made use of data obtained through the High Energy
494: Astrophysics Science Archive Research Center Online Service, provided by the
495: NASA/Goddard Space Flight Center. We also thank the {\it RXTE-ASM} and {\it
496: GINGA-ASM} teams for providing the valuable data. B. Paul was supported
497: by the Japan Society for the Promotion of Science through a fellowship.
498:
499: \end{acknowledgements}
500:
501: \begin{thebibliography}{}
502: %\begin{references}{}
503:
504: \bibitem{}
505: Byram, E. T., Chubb, T. A., \& Friedman, H. 1966, Science, 152, 66
506:
507: \bibitem{}
508: Corbet, R., Peele, A., \& Smith, D. A. 1996, IAUC, 6632
509:
510: \bibitem{}
511: Cowley, A. P., Crampton, D., \& Hutchings, J. B. 1979, ApJ, 231, 539
512:
513: \bibitem{}
514: Cowley, A. P., Crampton, D., Hutchings, J. B., Remillard, R., \& Penfold, J. E.
515: 1983, ApJ, 272, 118
516:
517: \bibitem{}
518: Cowley, A. P., et al. 1991, ApJ, 381, 526
519:
520: \bibitem{}
521: Dubus, G., Charles P. A., Long, K. S., \& Hakala, P. J. 1997, ApJ, 490, L47
522:
523: \bibitem{}
524: Ebisawa, K., Makino, F., Mitsuda, K., Belloni, T., Cowley, A. P., Schmidtke, P.
525: C., \& Treves, A. 1993, ApJ, 403, 684
526:
527: \bibitem{}
528: Elsner, R. F., Weisskopf, M. C., Darbro, W., Ramsey, B. D., Williams, A. C.,
529: Sutherland, P. G., \& Grindlay, J. E. 1986, ApJ, 308, 655
530:
531: \bibitem{}
532: Fabian, A. C., Eggleton, P. P., Pringle, J. E., \& Hut, P. 1986, ApJ, 305, 333
533:
534: \bibitem{}
535: Giacconi, R., Gursky, H., Kellogg, E., Levinson, R., Schreier, E., \&
536: Tananbaum, H. 1973, ApJ, 184, 227
537:
538: \bibitem{}
539: Hasinger, G., Langmeier, A., Sztajno, M., Truemper, J., \& Lewin, W. H. G.
540: 1986, Nature, 319, 469
541:
542: \bibitem{}
543: Hasinger, G., \& van der Klis, M. 1989, A\&A, 225, 79
544:
545: \bibitem{}
546: Holt, S. S., 1976, Ap\&SS, 42, 123
547:
548: \bibitem{}
549: Horne, J. H., \& Baliunas, S. L. 1986, ApJ, 302, 757
550:
551: \bibitem{}
552: Johnston, M. D., Bradt, H. V., Doxsey, R. E., Gursky, H., Schwartz, D. A.,
553: Schwarz, J., \& van Paradijs, J. 1978, ApJ, 225, L59
554:
555: \bibitem{}
556: Kahn, S. M., \& Grindlay, J. E. 1984, ApJ, 281, 826
557:
558: \bibitem{}
559: Katz, J. I. 1973, Nature, 246, 87
560:
561: \bibitem{}
562: Kitamoto, S., Egoshi, W., Miyamoto, S., Tsunemi, H., Ling, J. C., Wheaton, W.
563: A., \& Paul, B. 2000, ApJ (in press)
564:
565: \bibitem{}
566: Kitamoto, S., Tsunemi, H., Miyamoto, S., \& Roussel-Dupre, D. 1993, ApJ, 403,
567: 315
568:
569: \bibitem{}
570: Kong, A. K. H., Charles, P. A., \& Kuulkers, E. 1998, New Astronomy, 3, 301
571:
572: \bibitem{}
573: Kuulkers, E., van der Klis, M., \& Vaughan, B. A. 1996, A\&A, 311, 197
574:
575: \bibitem{}
576: Kuulkers, E., Wijnands, R., \& van der Klis, M. 1999, MNRAS, 308, 485
577:
578: \bibitem{}
579: Lang, F. L., et al. 1981, ApJ, 246, L21
580:
581: \bibitem{}
582: Leong, C., Kellogg, E., Gursky, H., Tananbaum, H., \& Giaconni, R. 1971, ApJ,
583: 170, L67
584:
585: \bibitem{}
586: Levine, A. M. 1998, in "The Active X-ray Sky; results from BeppoSAX and
587: Rossi-XTE", Nuclear Physics B, (Proc. Suppl.), 69/1-3, 196
588:
589: \bibitem{}
590: Levine, A. M., Bradt, H., Cui, W., Jernigan, J. G., Morgan, E. H., Remillard,
591: R., Shirey, R. E., \& Smith, D. A. 1996, ApJ, 469, L33
592:
593: \bibitem{}
594: Lomb, N. R. 1976, Ap\&SS, 39, 447
595:
596: \bibitem{}
597: Maloney, P. R., Begelman, M. C., \& Pringle, J. E. 1996, ApJ, 472, 582
598:
599: \bibitem{}
600: Margon, B. 1984, ARA\&A, 22, 507
601:
602: \bibitem{}
603: Margon, B., Grandi, S. A., Stone, R. P. S., \& Ford, H. C. 1979, ApJ, 233, L63
604:
605: \bibitem{}
606: Meyer, F., in "Radiation Hydrodynamics in Stars and Compact Objects',
607: Proceedings of IAU Colloq. 89, 1986, Eds. D. Mihalas \& Karl-Heinz
608: A. Winkler, Springer-Verlag, P. 249, 1986
609:
610: \bibitem{}
611: \"Ogelman, H., Kahabka, P., Pietsch, W., Tr\"umper, J., \& Voges, W. 1985,
612: SSRv, 40 3470
613:
614: \bibitem{}
615: Osaki, Y. 1985, A\&A, 144, 369
616:
617: \bibitem{}
618: Peele, A. G., \& White, N. E. 1996, IAUC, 6524
619:
620: \bibitem{}
621: Pietsch, W., Voges, W., Pakull, M., \& Staubert, R. 1985, SSRv, 40, 371
622:
623: \bibitem{}
624: Priedhorsky, W. C., \& Holt, S. S. 1987, SSRv, 45, 291
625:
626: \bibitem{}
627: Priedhorsky, W. C., \& Terrell, J. 1984, ApJ, 280, 661
628:
629: \bibitem{}
630: Priedhorsky, W. C., Terrell, J., \& Holt, S. S. 1983, ApJ, 270, 233
631:
632: \bibitem{}
633: Roberts, W. J. 1974, ApJ, 187, 575
634:
635: \bibitem{}
636: Schandl, S. 1996, A\&A, 307, 95
637:
638: \bibitem{}
639: Schandl, S., \& Meyer, F. 1994, A\&A, 289, 149
640:
641: \bibitem{}
642: Scargle, J. D. 1982, ApJ, 263, 835
643:
644: \bibitem{}
645: Smale, A. P. 1998, ApJ, 498, L141
646:
647: \bibitem{}
648: Smale, A. P., \& Lochner, J. C. 1992, ApJ, 395, 582
649:
650: \bibitem{}
651: Tananbaum, H., Gursky, H., Kellogg, E., Giacconi, R., \& Jones, C. 1972,
652: ApJ, 177, L5
653:
654: \bibitem{}
655: Treves, A., Belloni, T., Chiappetti, L., Maraschi, L., Stella, L., Tanzi, E.
656: G., \& van Der Klis, M. 1988, ApJ, 325, 119
657:
658: \bibitem{}
659: Tr\"umper, J., Kahabka, P., \"Ogelman, H., Pietsch, W., Voges, W. 1986,
660: ApJ, 300, L63
661:
662: \bibitem{}
663: Tsunemi, H., Kitamoto, S., Manabe, M., Miyamoto, S., Yamashita, K., \&
664: Nakagawa, M. 1989, PASJ, 41, 391
665:
666: \bibitem{}
667: Vrtilek, S. D., Swank, J. H., Kelley, R. L., \& Kahn, S. M. 1988, ApJ, 329, 276
668:
669: \bibitem{}
670: Weisskopf, M. C., Darbro, W. A., Elsner, R. F., Williams, A. C., Kahn, S. M.,
671: Grindlay, J. E., Naranan, S., \& Sutherland, P. G. 1983, ApJ, 274, L65
672:
673: \bibitem{}
674: White, N. E., \& Marshall, F. E. 1984, ApJ, 281, 354
675:
676: \bibitem{}
677: Wijers, R. A. M. J., \& Pringle, J. E.. 1999, MNRAS, 308, 207
678:
679: \bibitem{}
680: Wijnands, R. A. D., Kuulkers, E., \& Smale A. P. 1996, ApJ, 473, L45
681:
682: \bibitem{}
683: Wijnands, R., Homan, J., van Der Klis, M., Kuulkers, E., van Paradijs, J.,
684: Lewin, W. H. G., Lamb, F. K., Psaltis, D., \& Vaughan, B. 1998, ApJ, 493, L87
685:
686: \bibitem{}
687: Wijnands, R. A. D., van der Klis, M., Kuulkers, E., Asai, K., \& Hasinger, G.
688: 1997, A\&A, 323, 399
689:
690: \bibitem{}
691: Wilms, J., Nowak, M. A., Dove, J. B., Pottschmidt, K., Heindl,
692: W. A., Begelman, M. C., \& Staubert, R. 1999, in Highlights in
693: X-Ray Astronomy in Honour of Joachim Trümper's 65th
694: Birthday, ed. B. Aschenbach \& M. J. Freyberg (MPE Rep. 272;
695: Garching: MPE), in press
696:
697: \bibitem{}
698: Wojdowski, P., Clark, G. W., Levine, A. M., Woo, J. W., \& Zhang, S. N.
699: 1998, ApJ, 502, 253
700:
701: \end{thebibliography}
702: %\end{references}
703:
704: %\placetable{tbl-1}
705: \begin{deluxetable}{lccc}
706: \footnotesize
707: \tablenum{1}
708: \tablecaption{Different periods and their significances\label{tbl-1}}
709: \tablewidth{0pt}
710: \tablehead{
711: \colhead{Instrument}&\colhead{Duration}&\colhead{Periods (day)}
712: &\colhead{Periods in small segments} \nl
713: \colhead{(energy band)}&\colhead{(days)}&\colhead{(false alarm prob.)}
714: &\colhead{(false alarm prob.)} \nl
715: }
716:
717: \startdata
718: & & {\bf Cyg X-2} & \nl
719: \nl
720: RXTE-ASM & 1078 & 40.4 (1 E-20), 68.8 (7 E-25) & 39.5 (4 E-9), 70.1 (1 E-12) \nl
721: (1.5--12 keV) & & & 39.6 (9 E-6), 56.3 (7 E-8), 72.4 (3 E-7) \nl
722: & & & 41.8 (2 E-7), 56.8 (5 E-7), 73.7 (1 E-10) \nl
723: GINGA-ASM & 1645 & 53.7 (1 E-4), 61.3 (1 E-7) & \nl
724: (1--20 keV) & \nl
725: ARIEL-5 & 1963 & 41.3, 46.2, 53.9\tablenotemark{a}, 77.4 (1 E-16) & 45.7 (3 E-3), 55.3 (1 E-5), 78 (0.9)\tablenotemark{b} \nl
726: (3--6 keV) & & & 43.2 (0.015), 68.6 (4 E-5), 78 (0.8)\tablenotemark{b} \nl
727: & & & 41 (0.025), 53 (1 E-4), 78 (0.07)\tablenotemark{b} \nl
728: VELA-5B & 3675 & 33.9 (0.4), 77.4 (0.03) \nl
729: (3--12 keV) & \nl
730: \nl
731: & & {\bf LMC X-3} & \nl
732: \nl
733: RXTE-ASM & 1080 & 104 (3 E-38), 169 (2 E-27), 216 (1 E-28) &
734: 104 (3 E-27), 178 (2 E-35) \nl
735: (1.5--12 keV) & & & 102 (6 E-12), 222 (1 E-11) \nl
736: GINGA-ASM & 1682 & 105 (1.5 E-3), 214 (1 E-10), 328 (2 E-5) \nl
737: (1--20 keV) & \nl
738: HEAO 1 & 506 & 99 (2 E-16), 203 (4 E-12) \nl
739: (1--13 keV) \nl
740: ARIEL-5 & 1903 & 96 (8 E-3), 130 (0.03) \nl
741: (3--6 keV) \nl
742:
743:
744: \tablenotetext{a}{41.3 (1 E-15), 46.2 (7 E-20), 53.9 (2 E-23)}
745: \tablenotetext{b}{In the periodograms generated from the segments of
746: ARIEL 5 light curve, the peaks near 78 days are barely visible, but
747: in the periodogram of the complete light curve it is highly significant
748: (see Figure 2).}
749: \enddata
750: \end{deluxetable}
751:
752: \end{document}
753:
754: \figcaption[fig1.ps]{The {\it RXTE-ASM} (1.5--12 keV) and {\it GINGA-ASM}
755: (1--20 keV) light curves of Cyg X-2. The days of the observations are given
756: in truncated Julian days in the figure.}
757:
758: \figcaption[fig2.ps]{The Lomb-Scargle periodograms obtained from the
759: {\it RXTE-ASM, GINGA-ASM, ARIEL 5} and {\it VELA 5B} light curves of
760: Cyg X-2. The horizontal dashed line represent the 99\% confidence limits
761: (see text for details).}
762:
763: \figcaption[fig3.ps]{Three segments of the {\it RXTE-ASM} light curve of
764: length 358 days each and the corresponding Lomb-Scargle periodograms are
765: shown on the left and right hand sides respectively. The 99\% confidence
766: limits are shown with the dashed horizontal lines.}
767:
768: \figcaption[fig4.ps]{The long term light curves of LMC X-3 obtained with
769: the {\it RXTE-ASM, GINGA-ASM} and {\it HEAO~1} satellites. The energy ranges
770: are 1.5--12 keV, 1--20 keV and 1--13 keV respectively.}
771:
772: \figcaption[fig5.ps]{The Lomb-Scargle periodograms generated from the light
773: curves of LMC X-3 obtained with the {\it RXTE-ASM, GINGA-ASM, HEAO~1} and
774: {\it ARIEL 5} detectors. The dashed horizontal lines indicate the 99\%
775: confidence limits.}
776:
777: \figcaption[fig6.ps]{The Lomb-Scargle periodograms obtained from two segments
778: of the {\it RXTE-ASM} light curve of LMC X-3. Each segment of the light curve
779: is of 540 days duration. The dashed horizontal lines indicate the 99\%
780: confidence limits.}
781:
782: \figcaption[fig9.ps]{Hardness ratios HR1 (3-5 keV / 1.5-5 keV) and HR2
783: (5-12 keV/ 3-5 keV) of Cyg X-2 and LMC X-3 obtained from the {\it RXTE-ASM}
784: data are plotted against the total intensity. Cyg X-2 data points are averaged
785: for 3 days and LMC X-3 data points are averaged for 10 days.}
786:
787: