低質量X光雙星系統4U1820-30由一個中子星及約20%成簡併狀態之白矮星所組成，位於球狀星團NGC6624中心附近，除了685秒的軌道週期之外，另外存在著長達約170天的X-ray光變長週期，Chou & Griendary (2001)由其週期之穩定性，因而以階級式三星系統 (hierarchical triple system)來解釋，認為4U1820-30的這個長週期是由於第三顆星影響其內部軌道離心率變化所造成的。 本論文主要目的在研究4U1820-30約170天長週期的演變。我們使用RXTE上ASM從1996年到2009年的資料來分析，並使用RXTE/PCA及Swift/BAT的資料來作比較，觀察週期在不同能量區段上有何差異。所使用的是Lomb-Scargle periodogram的方法測量週期，並利用動態頻譜 (dynamic power spectra)的方式觀察不同時期所對應的週期與強度變化。另一方面，利用O-C分析的方法得到1996 ~ 2009年期間的相位演化。 然而，以Lomb-Scargle方法所測出的週期，顯示4U1820-30的X-ray光變長週期在TJD13000後減少至約168天，而以O-C分析方法所得到的相位演化作星曆表修正後，不論是以突然的週期改變(約2%)或是此週期持續性地在變化，都無法合理地用階級性三星系統中的雙星或是其第三顆星週期變化來解釋，使得我們對於4U1820-30約170天的長週期X-ray光變是否真正的是由階級性三星系統所產生的提出質疑。另一方面，在動態頻譜中TJD1100 ~ 12300之間出現約170天的週期訊號變弱，我們認為這是由於second minima在這段時期的增加所造成，但即使不考慮這段週期不明顯的時期，在相位演化圖上仍然呈現相位不連續的情形，這樣的現象或許也暗示了週期在TJD11000 ~ 12300之間發生了異常的變化。 此外，藉由不同儀器所分析出的資料顯示，以RXTE/ASM及PCA所得到的結果是一致的，而Swift/BAT所測得到週期明顯較短，雖然Farrell et al. (2008)曾提出4U1820-30約170天的週期只會出現在24keV以下，但這仍然無法解釋為何Swift/BAT所測得的週期會不一樣。本研究裡還檢查了X-ray爆發時所對應的光變曲線，顯示不論週期發生改變前後，X-ray爆發發生的時間仍然在low state附近 ±22天內，再一次確認此光度變化是由於吸積率所造成的。 4U1820-30 is a low mass X-ray binary (LMXB) system, located near the center of the globular cluster NGC 6624, composed of a neutron star and a ~20% degeneracy white dwarf. Besides its 685s orbital period, it also has a ~ 170d long-term variation in X-ray band. Chou & Grindlay (2001) proposed the hierarchical triple model for 4U1820-30 system based on the stability of ~171d periodicity. The purpose of this thesis is to investigate the long-term modulation of this source. The X-ray light curve adopted for this study was from the observations of All-Sky Monitor (ASM) on board Rossi X-ray Timing Explorer (RXTE) from 1996 to 2009. We also used the data from Proportional Counter Array (PCA) on RXTE and Swift Burst Alert Telescope (BAT) to compare the results between different energy bands. The Lomb-Scargle periodogram was applied to verify the periodicity of long-term modulation. We also performed dynamic power spectrum analysis to trace the modulation power variation, as well as the period change. On the other hand, phase evolution evaluated by Observed – Calculated (O-C) method was also included to study long-term modulation of 4U1820-30. However, this period is likely shift to ~ 168d from it power spectrum after 2004. The phase evolution also shows negative drift after 2004. But neither of the abrupt period change( ~ 2%) nor the period derivative caused by binary or third companion’s orbital period, can be explained by triple model. These evidences above all imply that the period of long-term modulation of 4U1820-30 is not as stable as the one proposed by Chou & Grindlay (2001), which is a challenge of the triple model for this system. Furthermore, ~ 170d power between TJD 11000 and TJD 12300 became weaker in the dynamic power spectrum, due to enhancements of secondary minima. Nevertheless, ~ 167d modulation from phase evolution shows unreasonable discontinuous phase during weak power interval, which is also unlikely to be explained with triple model. The period detected from RXTE/ASM and PCA is consistent, but the data from Swift/BAT show period shift to ~160d after 2004. Although Farrell et al. (2008) proposed that ~ 170d period of 4U1820-30 could be detected only up to 24 keV, it still cannot explain why the period shifts. In our studies, we also checked the time of occurrence of X-ray bursts. All the X-ray bursts detected within ±22d of the intensity minima rather than predicted minima proposed by Chou & Grindlay (2001), which indicates that the long-term modulation is a change in the accretion rate.
