Variability of the symbiotic X-ray binary GX 1+4: Enhanced activity near periastron passage
AAstronomy & Astrophysics manuscript no. GX1 + (cid:13) ESO 2018October 28, 2018
Variability of the symbiotic X-ray binary GX 1+4:
Enhanced activity near periastron passage
Krystian Iłkiewicz , Joanna Mikołajewska , and Berto Monard Nicolaus Copernicus Astronomical Center, Polish Academy of Sciences, Bartycka 18, 00716 Warsaw, Poland Kleinkaroo Observatory, Calitzdorp, Western Cape, South AfricaReceived ... / Accepted ...
ABSTRACT
Context.
GX 1 + Aims.
We study variability of GX 1 + Methods.
The presented X-ray observations are from INTEGRAL Soft Gamma-Ray Imager and RXTE All Sky Monitor. The opticalobservations are from INTEGRAL Optical Monitoring Camera.
Results.
The variability of GX 1 + >
17 keV) is dominated by ∼ Key words.
X-rays: binaries – binaries: symbiotic – stars: individual: GX 1 +
1. Introduction
GX 1 + = V2116 Oph is an X-ray binary that was one of thebrightest X-ray objects known at time of discovery (Lewin et al.1971). It is harboring a pulsar with a spin period of ∼ +
4. The object showed spectralfeatures typical for a symbiotic star (SySt), i.e. a binary systemwith a red giant (RG) donor (Davidsen et al. 1977). The symbi-otic nature of GX 1 + + > ∼ > ∼ ∼ α emission line and concluded that the orbital period has to be Send o ff print requests to : K. Iłkiewicz, e-mail: [email protected] significantly longer that the ∼ I light curve made by the OGLE IV project, andfound a quasi-periodic variations with a timescale of 50–75 days,which they have attributed to changes in the accretion rate. Theyalso found a strong peak in the power spectrum corresponding tothe period of 295 ± ff ected by substantial seasonal gaps intheir light curve.Finally, Hinkle et al. (2006), based on high-resolution near-infrared spectroscopy spanning nearly 5 yr, determined the firstspectroscopic orbit of the red giant. They found an orbital pe-riod of 1161 ±
12d and modest eccentricity e = ± + f ( m ) = . ± .
026 is con-sistent with the secondary being a neutron star (NS), and the redgiant is the less massive component ( < ∼ . (cid:12) ). González-Galánet al. (2012) showed that the spin-up of the NS is marginally cor-related with periastron passage times calculated by Hinkle et al.(2006).Here we report analysis of GX 1 +
2. Observations
We carried out optical observations of GX 1 + Ic filter. Each observation was the result of sev- Article number, page 1 of 8 a r X i v : . [ a s t r o - ph . H E ] F e b & A proofs: manuscript no. GX1 + Fig. 1.
Light curves of GX 1 + ff erent spectral ranges. The green areas indicate time of periastron passage (Eq. 1) and the gray areas indicatetime of the potential eclipse of the neutron star (Eq. 2). eral (3 to 6) individual exposures, which were calibrated (dark-subtraction and flatfielding) and stacked selectively. Magnitudeswere derived from di ff erential photometry to nearby referencestars using the single image mode of AIP4 image processingsoftware. The photometric precision of the derived Ic magni- tudes is estimated to be better than 0.1 mag. Moreover, we sup-plied our photometry with observations in the I -band scannedfrom Majczyna et al. (2015). We shifted our observations sothat the mean flux during the period covered both by ours andMajczyna et al. (2015) observations would be equal. This was Article number, page 2 of 8rystian Iłkiewicz et al.: Variability of the symbiotic X-ray binary GX 1 + done in order to ensure that the same photometric zeropoint wasadapted in both datasets. Afterwards, we transformed the mag-nitudes to flux density using calibration of Bessell (1979). Theobservations are presented in Table B.1, Figs. 2 and 2.We gathered optical observations in V filter from Interna-tional Gamma Ray Astrophysics Laboratory (INTEGRAL) Op-tical Monitoring Camera (OMC; Mas-Hesse et al. 2003), a re-fractive telescope with a 50 mm aperture. The telescope isequipped with a Johnson V filter, has 1024x1024 pixels imag-ing area CCD and 5 ◦ x5 ◦ field of view. GX 1 + <
3. The ASMdata was partially analyzed by Corbet et al. (2008), who did notfind evidence of variability related to the orbital motion.Hard X-ray data is from the INTEGRAL Soft Gamma-RayImager (ISGRI; Lebrun et al. 2003). This is a low energy detectorof the Imager on Board the INTEGRAL Satellite (IBIS; Uber-tini et al. 2003). The employed observations were from spectralranges 17.3–22.1 keV, 22.1–30.0 keV and 30.0–63.3 keV. Theobservations are binned with a bin size of 1d. Data points withsignal to noise ratio < (HEAVENS; Walter et al. 2010). The lightcurves are presented in Figs 1 and 2.
3. Results
Fourier analysis of the observations was performed using the dis-crete Fourier transform method in the
Period
04 program (Lenz& Breger 2005). The resulting power spectra are presented inFig. 3. The power spectra are apparently dominated by a low-frequency noise. In order to estimate the significance level ofpeaks in the power spectrum we assumed the noise level to beequal to an average amplitude in the interval ± − to agiven frequency. We constructed significance curve adopting sig-nal to noise ratio > / heavens light curves in the 17.3-63.3 keV range (Fig. 1) consists of se-ries of flares with characteristic timescales of 50–70d similar tothe quasi-periodic changes reported by Majczyna et al. (2015),and there is no permanent X-ray emission. The comparison ofthe ISGRI light and I light curves (Fig. 1,2) shows that virtuallythe same phenomenon is observed in the hard X-rays and theoptical I light. The quasi-periodic nature of flares caused a lowfrequency noise in the power spectra, because of which the or-bital period was not formally detected. However, the flares seemto be particularly prominent when the binary approaches peri-astron. Majczyna et al. (2015) suggested that optical brighten-ing with low and high amplitude could be caused by a di ff erentmechanism. Using the hard X-ray observations ( > T = JD , , ± ± . ± . × E (1)for the periastron passage and ephemeris T = JD , , ± ± . ± . × E (2)for the potential eclipse of the neutron star (Hinkle et al. 2006).Moreover no brightenings in X-rays are observed around thetime of possible NS eclipse, which leads to skewed phase plotof hard X-ray observations (Fig. 5). The variations in amplitudeof quasi-periodic changes confirms the orbital period derived byHinkle et al. (2006) and indicates that the big and low amplitudebrightenings are the same phenomenon with di ff erent amplitudeat di ff erent orbital phase. We note that one of the flares in I bandwas observed near the time of possible NS eclipse (Fig. 2), how-ever the uncertainty of the time of potential eclipse is larger thanthe timescale of the flare, because of accumulation of errors inephemeris of Hinkle et al. (2006) over few cycles. In fact, in-creasing the orbital period by ∼ σ and / or somewhat later pe-riastron passage would remove the problem with appearance offlares during the eclipse. Alternatively if flares during the NSeclipse would be observed in the optical light only and not inX-rays this could indicate that the source of the optical light is atlarge distance from the NS (e.g. in jets), and it is not eclipsed.Enhanced mass transfer rate at periastron as a source of big-ger amplitude of the flares is consistent with the fact that themaximum flux is observed at orbital phase φ ∼ . φ = I ) range (see Fig. 2; Majczyna et al. 2015). Similar cor-relation in GX 1 + α emissionline (Sood et al. 1995; Greenhill et al. 1995). It is worth notingthat decline of both H α and hard X-ray emission from the July–September 1993 flare coincided to within a few days, howeverthe flare in H α started ∼
30d before the flare in X-rays. Greenhillet al. (1995) noticed that this points to the fact that both X-rays
Article number, page 3 of 8 & A proofs: manuscript no. GX1 + Fig. 2.
Example of variability on short time-scale in 30.0–63.3 keV range (black points; left axis) together with observations in the I -band (redpoints; right axis) scanned from Majczyna et al. (2015). The green area indicates time of periastron passage (Eq. 1) and the gray areas indicatetime of the potential eclipse of the neutron star (Eq. 2). and H α variability has the same source, however H α variabilitycould not be directly caused by changes in X-ray radiation. If thebrightening is caused by a short increase in the mass accretionrate trough the accretion disc, the time lag can be estimated usingmass of the NS, M NS , mass accretion rate, ˙ M , viscosity, α , andthe e ff ective temperature corresponding to the maximum in theH α emission, T e ff , H α (Bisnovatyi-Kogan & Giovannelli 2016).After adopting the total X-ray luminosity L X (cid:39) × erg s − (e.g. Galloway et al. 2001) and assuming the typical parametersof the NS, M NS = (cid:12) (Thorsett & Chakrabarty 1999) and ra-dius R NS =
10 km, the mass accretion rate can be estimated fromthe equation ˙ M = L X R NS / ( G M NS ) to be ∼ × − M (cid:12) yr − .Assuming T e ff , H α = α = α emission to be ∼ I and X-ray flares, hence, in contrary to theH α emission, the optical I probably has a di ff erent origin thanthe increased mass transfer rate trough the outer parts of accre-tion disc. Although the timescales of the semi-periodic I -bandvariability is similar to those of semi-regular (SR) pulsationsof red giants, such SR pulsations have much lower amplitudes, < ∼ . ∆ I up to 1 mag (Fig. 2; Majczynaet al. 2015), and their periods are well defined (e.g. Gromadzkiet al. 2013). Moreover, this variability practically vanishes dur-ing inferior conjuctions when the red giant should dominate the I -band light. On the other hand, similar flares on time scaleof tens of days have been observed simultaneously in X-rays,optical and near infrared emission in some accreting low-massX-ray binaries (LMXB; see e.g Veledina et al. 2013, and ref-erences therein). While they seem to be due to unstable disk-accretion, the nature of the optical and near infrared emission re-mains poorly understood. In particular, it has been suggested thatthese bands are dominated by either the jet emission, extendedhot accretion flow or outer irradiated part of the accretion diskheated by the X-rays. Although most of the LMXB discussed by Veledina et al. (2013) host black holes, their accretion diskbehavior should not be much di ff erent from those with neutronstars. It is thus tempting to assume similar origin of the flaredemission in GX 1 +
4; all these scenarios would produce opticalradiation with much shorter delay time compared to the X-rays,which would not be detected in our observations. It is also worthto note that the X-ray and optical variability has similar relativeamplitude (Fig. 2), which in future could be used to determinethe origin of the optical radiation (Russell et al. 2006).The sawtooth-wave like shape of the binned phase plot of ob-servations in 17.3-63.3 keV range (Fig. 5) is somewhat unusualfor an X-ray binary, given that it shows slow rise and fast decay,while the opposite is usually observed, i.e. fast rise, and slow de-cay. Another object with sawtooth-wave like variability skewedin the same way was X-ray nova GRS 1716-249 (Hjellming et al.1996). While variability of this object was not related to orbitalmotion, the sawtooth-wave was almost identical to the binnedphase plot of observations in 17.3-63.3 keV range. Moreover,it was observed in similar spectral range, namely 20–100 keV.Hjellming et al. (1996) suggested, based on their radio observa-tions, that variability of GRS 1716-249 was related to ejectionof relativistic particles in a jet. Esin et al. (1998) proposed, thatGRS 1716-249 during the flare maxima was in the the low spec-tral state and the rapid decay was related to rapid increase in themass accretion rate that precipitated a corresponding decrease intransition radius. During minima in the 20–100 keV range, GRS1716-249 would be in high energy state, and the transition ra-dius would be close to the last stable orbit. Similar mechanismmight take place in GX 1 +
4, given that the fast decline on binnedphase-plot fallows shortly after the periastron passage (Fig. 5),when increased mass transfer rate is expected. In that case, therapid decline would only coincidentally be observed close to thetime of spectroscopic conjunction.The binned phase-plot is a measure of amplitude of the ∼ Article number, page 4 of 8rystian Iłkiewicz et al.: Variability of the symbiotic X-ray binary GX 1 + Fig. 3.
Power spectrum of X-ray observations of GX 1 + seems to be the lowest near the observed minimum (Fig 5). Themore likely explanation is that the quasi-periodic flares originatein a region in accretion disc that would be a ff ected by change inthe transition radius. Moreover, it seems that the sawtooth-waveshape is not always present. In the lightcurve in 17.3-63.3 keVrange the drop to minimum in years 2006-2007 is slower, resem-bling more a skewed sinusoidal function (Fig. 1). In the otherobserved drops in amplitude of the ∼ + V filter (Fig. 1, 5). The vari-ability may be combination of orbital period (Fig. 5) and shortterm variability discussed by Majczyna et al. (2015), but qualityof the data is not su ffi cient for any meaningful analysis. Howeverthe orbital period of 1160.8d is not excluded on basis of observa-tions in V filter, that have longer baseline than the observationsin I filter, which in turn show variability related to orbital motion(Fig. 5).The light-curve in 1.3–12.2 keV band (Fig. 1) shows irregu-lar variability. This is consistent with the fact that Corbet et al.(2008) did not detect orbital period on basis of the same obser-vations, although with shorter baseline. The lack of detectableorbital variability may be due to the fact that the variability inthis range is dominated by variable column density rather thanvariable mass transfer rate, as was observed e.g. in RT Cru and Fig. 4.
Top five panels: phase plot of GX 1 + ff erent spectral rangeswith the orbital period of 1160.8d (Hinkle et al. 2006), where orbitalphase φ = +
4. The hardness isdefined as a di ff erence between count rate in 30.0–63.3 keV range andin 17.3–22.1 keV range. Article number, page 5 of 8 & A proofs: manuscript no. GX1 + Fig. 5.
Same as Fig. 4, but with binned data points. The top panelshas been created with bin size of 0.2 of orbital phase. In case of otherpanels bin size of 0.1 of orbital phase was employed. The red errorbars represent standard error of mean. The black error bars representstandard deviation. suggested for T CrB (Kennea et al. 2009; Iłkiewicz et al. 2016).It is interesting to note that near the orbital phase φ ∼ .
4. Conclusions
In this work we analyzed long term X-ray and optical variabilityof a SyXB GX 1 +
4. The main conclusions are: – There is not permanent hard X-ray ( >
17 keV) emission inGX 1 +
4. Both hard X-ray and optical I band are dominatedby flares occurring quasi-periodically on ∼ α emission has had been observed. – The observations in the hard X-rays and optical I band con-firm the orbital period of 1160.8d derived on basis of IRspectroscopy by Hinkle et al. (2006). In particular, we seepronounced, regular behavior in hard X-rays and optical I band, while it is not visible in soft X-rays (Fig. 5). – The period of ∼ – The variability in the soft X-ray band shows no apparentperiodicity, but seems to be always at minimum during themaximum of hard X-rays. – Detection of orbital period in hard X-rays and no-detection insoft X-rays shows that the former may be better to search fororbital period in other SyXB. This would explain why orbitalperiod was not detected in 4U 1700 +
24 and 4U 1954 +
31 onbasis of observations in softer X-rays (Corbet et al. 2008).GX 1 + Acknowledgements.
We are grateful to Andrzej Zdziarski for a helpful dis-cussion. KI has been financed by the Polish Ministry of Science and HigherEducation Diamond Grant Programme via grant 0136 / DIA / /
43. Thisstudy has been partially financed by Polish National Science Centre grants2012 / / M / ST9 / / / A / ST9 / References
Alfonso-Garzón, J., Domingo, A., Mas-Hesse, J. M., & Giménez, A. 2012,A&A, 548, A79Arnaud, K. A. 1996, in Astronomical Society of the Pacific Conference Series,Vol. 101, Astronomical Data Analysis Software and Systems V, ed. G. H.Jacoby & J. Barnes, 17Bessell, M. S. 1979, PASP, 91, 589Bisnovatyi-Kogan, G. S. & Giovannelli, F. 2016, ArXiv e-prints[ arXiv:1605.07013 ]Breger, M., Stich, J., Garrido, R., et al. 1993, A&A, 271, 482
Article number, page 6 of 8rystian Iłkiewicz et al.: Variability of the symbiotic X-ray binary GX 1 + Chakrabarty, D. & Roche, P. 1997, ApJ, 489, 254Corbet, R. H. D., Sokoloski, J. L., Mukai, K., Markwardt, C. B., & Tueller, J.2008, ApJ, 675, 1424Cutler, E. P., Dennis, B. R., & Dolan, J. F. 1986, ApJ, 300, 551Davidsen, A., Malina, R., & Bowyer, S. 1977, ApJ, 211, 866Doty, J. P., Lewin, W. H. G., & Ho ff man, J. A. 1981, ApJ, 243, 257Esin, A. A., Narayan, R., Cui, W., Grove, J. E., & Zhang, S.-N. 1998, ApJ, 505,854Ferrigno, C., Segreto, A., Santangelo, A., et al. 2007, A&A, 462, 995Galloway, D. K., Giles, A. B., Wu, K., & Greenhill, J. G. 2001, MNRAS, 325,419Glass, I. S. & Feast, M. W. 1973, Nature Physical Science, 245, 39González-Galán, A., Kuulkers, E., Kretschmar, P., et al. 2012, A&A, 537, A66Greenhill, J. G., Watson, R. D., Tobin, W., Pritchard, J. D., & Clark, M. 1995,MNRAS, 274, L59Gromadzki, M., Mikołajewska, J., & Soszy´nski, I. 2013, Acta Astron., 63, 405Hinkle, K. H., Fekel, F. C., Joyce, R. R., et al. 2006, ApJ, 641, 479Hjellming, R. M., Rupen, M. P., Shrader, C. R., et al. 1996, ApJ, 470, L105Iłkiewicz, K., Mikołajewska, J., Stoyanov, K., Manousakis, A., & Miszalski, B.2016, MNRAS, 462, 2695Jablonski, F. J., Pereira, M. G., Braga, J., & Gneiding, C. D. 1997, ApJ, 482,L171Kennea, J. A., Mukai, K., Sokoloski, J. L., et al. 2009, ApJ, 701, 1992Lebrun, F., Leray, J. P., Lavocat, P., et al. 2003, A&A, 411, L141Lenz, P. & Breger, M. 2005, Communications in Asteroseismology, 146, 53Lewin, W. H. G., Ricker, G. R., & McClintock, J. E. 1971, ApJ, 169, L17Majczyna, A., Madej, J., Nalezyty, M., Rozanska, A., & Udalski, A. 2015, inProceedings of the Polish Astronomical Society, Vol. 3, XXXVII Polish As-tronomical Society Meeting, ed. A. Rozanska & M. Bejger, 133Mas-Hesse, J. M., Giménez, A., Culhane, J. L., et al. 2003, A&A, 411, L261Masetti, N., Orlandini, M., Palazzi, E., Amati, L., & Frontera, F. 2006, A&A,453, 295Nagase, F. 1989, PASJ, 41, 1Pereira, M. G., Braga, J., & Jablonski, F. 1999, ApJ, 526, L105Russell, D. M., Fender, R. P., Hynes, R. I., et al. 2006, MNRAS, 371, 1334Sood, R. K., James, S. D., Lawson, W. A., et al. 1995, Advances in Space Re-search, 16, 131Thorsett, S. E. & Chakrabarty, D. 1999, ApJ, 512, 288Ubertini, P., Lebrun, F., Di Cocco, G., et al. 2003, A&A, 411, L131Veledina, A., Poutanen, J., & Vurm, I. 2013, MNRAS, 430, 3196Walter, R., Rohlfs, R., Meharga, M. T., et al. 2010, in Eighth Integral Workshop.The Restless Gamma-ray Universe (INTEGRAL 2010), 162Zdziarski, A. A., Segreto, A., & Pooley, G. G. 2016, MNRAS, 456, 775 Appendix A: The 304d period
In this section we perform similar analysis as in the main bodyof the text, but for ∼ ∼ > ∼ ∼ Appendix B: Observational data
Table B.1.
Observations of GX 1 + MJD Ic [mag] Flux density × [ erg s − cm − Hz − ]56339.00 15.02 2.5156355.00 15.09 2.3556368.00 14.91 2.7856386.00 14.68 3.4356402.00 14.71 3.3456414.00 14.69 3.4056430.00 14.70 3.3756440.80 14.57 3.8056448.80 14.43 4.3256472.70 14.72 3.3156490.80 14.45 4.2556502.80 14.28 4.9656509.80 14.48 4.1356517.80 14.67 3.4756527.80 14.67 3.4756542.80 14.50 4.0556561.80 14.35 4.6556573.80 14.37 4.5756692.10 14.43 4.3256710.10 14.62 3.6356722.00 14.95 2.6856738.00 14.96 2.6556743.90 15.00 2.5656752.90 14.82 3.0256756.10 14.68 3.4356772.00 14.49 4.0956786.90 14.77 3.1656806.00 14.47 4.1756806.90 14.27 5.0156825.05 14.47 4.1756845.90 14.87 2.8856867.90 15.02 2.5156884.90 14.98 2.6156892.85 14.80 3.0856900.85 14.87 2.8856921.80 14.80 3.0856943.80 15.07 2.4057070.10 14.67 3.4757085.10 14.65 3.5357100.05 14.83 2.9957111.15 14.86 2.9157136.10 14.84 2.9657162.00 14.96 2.6557187.95 14.66 3.5057217.80 14.82 3.0257226.75 14.93 2.7357240.75 14.83 2.9957255.80 14.78 3.1357299.70 14.69 3.4057323.75 14.83 2.9957418.10 14.60 3.7057428.10 14.67 3.4757445.10 14.65 3.5357478.15 14.38 4.5357481.05 14.60 3.7057508.95 14.73 3.2857534.84 14.90 2.8057567.93 14.59 3.7357601.82 14.52 3.9857642.82 14.39 4.4957686.78 14.51 4.0257782.12 13.97 6.6157785.10 13.97 6.6157790.12 14.09 5.91 Article number, page 7 of 8 & A proofs: manuscript no. GX1 + Fig. A.1.
Same as Fig. 4, but with P = Fig. A.2.
Same as Fig. 5, but with P ==