4U 1820-30 為一坐落於球狀星團 NGC6624 中心附近的低質量 X 光雙星系統，該系統由一中子星及白矮星組成，雙星軌道週期為 685 秒，除了雙星軌道週期外，另外存在一約 170 天的 X-光超軌道光變週期，Chou & Grindlay (2001)計算出此超軌道光變週期為 171.033 天以及週期變化率上限 P˙/P < 2.2 × 10−4 yr^−1 認為該週期是穩定的，以階級性三星系統來解釋其週期穩定性。 本論文之主要研究方向為研究該系統的 170 天的超軌道光變週期穩定性，利用 Ginga、RXTE、Swift、MAXI 以及 Fermi 等不同時期的 X 光望遠鏡的觀測數據來分析，資料的時間跨度從 1987 年至 2024 年，共 37 年。 功率譜分析的結果顯示，在這 37 年間，4U 1820-30 的超軌道光變週期從 171 天下降至 166 天，根據近一步的相位分析，本研究提出超軌道光變週期兩種可能的變化模型: 突變模型與二次曲線模型，突變模型顯示 4U 1820-30 的超軌道光變週期可能在MJD52502.83 ± 609.55 突然變化，由原先的 170.78 ± 0.79 天變化至 167.46 ± 0.18 天；而二次曲線模型則假設 4U 1820-30 的超軌道光變週期為連續變化，所得出的週期變化率為P˙ /P = (−8.41±1.58) ×10−4 yr^−1，遠大於Chou & Grindlay (2001)計算出的週期變化率上限。 種種跡證說明 4U 1820-30 的超軌道光變週期遠不如三星模型想像中的穩定，嚴重的挑戰了階級性三星模型，根據Zdziarski et al. (2007b)的發現，685 秒軌道光變週期振輻與相位與吸積率間存在相關性，這說明超軌道光變來源於伴星質量損失率變化，本研究使用RXTE、NICER 的觀測數據再次驗證了這項關係，根據以上結果，本研究提出輻照不穩定假說來解釋超軌道光變的變化。 ;4U 1820-30 is a low-mass X-ray binary system located near the center of the globular cluster NGC6624. This system is composed of a neutron star and a white dwarf with an orbital period of ~685 s. Except for this orbital period, 4U 1820-30 also exhibits a superorbital period of ~170 days. Chou & Grindlay (2001) observed a 171.033-day superorbital period and a period derivative upper limit P˙/P < 2.2 × 10−4 yr^−1. They concluded this period is stable and explained it by a hierarchical triple system. This research is to verify the stability of the 170-day superorbital period of 4U 1820-30. We collected the light curves observed from different eras, by the instruments: Ginga, RXTE, Swift, MAXI, and Fermi, total time span of 37 years, from 1987 to 2024. From power spectral analysis, we discovered that the superorbital modulation of 4U 1820-30 decreased from 171 days to 166 days during these 37 years. The phase analysis revealed that the superorbital phase evolution can be described by the glitch and the quadratic models. The glitch model implies that the superorbital could suddenly change on MJD52502.83 ± 609.55 from 170.78 ± 0.79 days to 167.46 ± 0.18 days. The quadratic model assumed that the superorbital period changes smoothly. A significant period derivative evaluated from phase evolution is P˙ /P = (−8.41 ± 1.58) × 10−4 yr^−1, inconsistent with the result from Chou & Grindlay (2001). We conclude that the superorbital period of 4U 1820-30 is not as stable as the prediction of the triple model, which strongly challenges this model. On the other hand, according to the discovery of Zdziarski et al. (2007b), the orbital modulation is correlated to the accretion rate. This implies that the superorbital modulation is induced by the mass-loss rate variation from the companion. We further analyzed the RXTE and NICER data and confirmed Zdziarski et al. (2007b) result. Therefore, we proposed the irradiation-induced instability scenario to explain the variation of the superorbital modulation period.
