Swift BAT (2005-2026) and MAXI (2009-2026) light curves show no new X-ray period from 2.1 to 1000 days that both monitors confirm; a 72.3-day ISS comb affects about half of MAXI sources
We searched the public daily light curves of the Swift BAT transient monitor (1,055 sources, 15-50 keV, 2005-2026) and of MAXI GSC (452 sources, 2-20 keV, 2009-2026) for periods from 2.1 to 1000 days. We used a semi-weighted Lomb-Scargle periodogram with a running-median red-noise normalisation and a 4-harmonic statistic. We calibrated it with 106 BAT blank-sky positions (none above S = 6.6), shuffled source pairs and 8,640 injected sinusoids. Three independent tests were applied: BAT x MAXI coincidence for 360 sources (368 confirmed peaks, 1.1 per cent false rate), BAT split-half persistence, and a BAT out-of-sample test (discovery in one decade, prediction of period and phase in the other; 170 of 431 peaks confirmed, 0.5 per cent false rate). All confirmed periods are catalogued orbital or superorbital periods, their harmonics, aliases, window side peaks or orbit-superorbit sidebands in 28-42 sources. The recurrence of MAXI J0903-531 (59.6-59.8 d in BAT 2021-2026, against a catalogued 57-d orbit) was already reported. A list of 15 weak BAT-only candidates, for example IGR J17200-3116 at 12.57 d, did not pass the out-of-sample test. For quiet sources, a sinusoid with a semi-amplitude of 2 mCrab at 2.2-100 d is recovered in 86-91 per cent of cases. Recovery is much lower for long periods and strongly variable sources. MAXI light curves contain a 72.3-day comb from the nodal precession of the ISS: 45-46 per cent of all MAXI sources have S1 > 3 at 72.3 d or 36.2 d, and harmonics are significant down to 3.46 d. MAXI also shows ISS sidebands of real periods (Vela X-1 at 7.98 d, Cen X-3 at 2.16 d) and orbits of neighbours up to 1.7 deg away (4U 1700-37 in GX 349+2). BAT shows no common line. Limitations: periods below 2.1 d were not searched, the red-noise model is approximate, quasi-periodic signals lose significance, and source identification depends on catalogues and a short literature list.
X-ray binaries · orbital periods · superorbital periods · Swift BAT · MAXI · periodogram · instrumental artifacts · ISS precession · source confusion · null result
References
14- usesdoi:10.1088/0067-0049/209/1/14
Swift BAT transient monitor light curves (Krimm et al. 2013).
- usesdoi:10.1093/pasj/61.5.999
MAXI mission and GSC light curves (Matsuoka et al. 2009).
- useshttps://swift.gsfc.nasa.gov/results/transients/
Source of the 1,162 BAT daily light curves, downloaded 2026-11-09.
- useshttp://maxi.riken.jp/top/slist.html
Source list and 457 MAXI daily light curves, downloaded 2026-11-09.
- usesdoi:10.1051/0004-6361/202245728
Galactic HMXB catalogue (Neumann et al. 2023) for known periods.
- useshttps://vizier.nao.ac.jp/viz-bin/VizieR?-source=J/A%2BA/675/A199
Galactic LMXB catalogue (Avakyan et al. 2023) for known periods.
- useshttps://vizier.nao.ac.jp/viz-bin/VizieR?-source=J/A%2BA/671/A149
HMXB catalogue with superorbital periods (Fortin et al. 2023).
- usesdoi:10.1051/0004-6361:20030330
Ritter and Kolb catalogue (B/cb) orbital periods.
- supportsdoi:10.1088/0004-637X/778/1/45
Superorbital periods of wind-fed HMXBs; we confirm those in 4U 1909+07 and others in later BAT and MAXI data.
- supportsdoi:10.48550/arXiv.1504.01105
Reported the ~70-d MAXI precession artifact for gamma Cas and nearby sources; we quantify it over all MAXI sources.
- doi:10.48550/arXiv.2206.01505
MAXI overview that describes the 72-d scan precession.
- challengeshttps://www.astronomerstelegram.org/?read=14730
Proposed a ~57-d orbit for MAXI J0903-531; the BAT recurrence in 2021-2026 is 59.6-59.8 d.
- supportshttps://baas.aas.org/pub/2023n4i116p14
HEAD 2023 abstract that reported a ~59-d outburst recurrence of MAXI J0903-531.
- https://www.astronomerstelegram.org/?read=16787
Warns that MAXI precession lines make blind MAXI period searches unreliable.
Cited by
1 publicationPublic catalog entries on Sciscape.
- reviewsSix known multiple-trigger bursts account for the one-day, two-degree GBM pair excess in 4418 selected triggers
Campaign context for independent instrument checks. No code or derived results are reused.