本篇論文將探討吸積毫秒脈衝星(accreting millisecond X-ray pulsars, AMXP) XTE J1807-294於2003年爆發對時間噪音的分析結果。XTE J1807-294在先前的研究發現具有異常的相位變化，且與X光強度有關，使得我們無法準確得到參數，Chou et al. (2008) 指出中子星熱點移動是造成相位異常的主要因素，後面稱之為與X光強度有關的時間噪音(flux-dependent timing noise)。如何有效地消除XTE J1807-294時間噪音以及得到更好的參數是本研究主要的目標。 一開始我們使用常規的脈衝抵達時間延遲技術去求得XTE J1807-294的自轉參數，然而求得的χ^2 值1055.55 (d.o.f. = 74) 太大，由此可以看出時間噪音對於XTE J1807-294有明顯的影響，使用單一參數方法對求得XTE J1807-294精確參數效果有限求得的χ^2 值1012.79 (d.o.f. = 72)，發現對參數改善的效果有限，因此我們藉由X光強度衰減趨勢，將整筆資料分為三個部份，並取得χ^2 值為312.08 (d.o.f. = 72)，相較於消除時間噪音前有明顯的改善。結果表明模型可以有效的消除時間噪音，得到的自轉參數也更接近真實數值，然而無法單一參數直接消除整筆資料的時間噪音也顯示模型有改善的空間。最後我們發現消除時間噪音後結果顯示中子星自轉頻率下降，而非先前研究呈現自轉頻率上升。 ;We present our analysis results in the study of flux-dependent timing noise for the accreting millisecond X-ray pulsars (AMXP) XTE J1807-294 during its 2003 outburst. XTE J1807-294 was found to have abnormal phase drift after being discovered shortly. Due to abnormal phase drift, which relates to X-ray flux, it is hard to get precise parameters of XTE J1807-294. Chou et al. (2008) suggested that anomalous phase shift is cause by hot spot motion, which is known as flux-dependent timing noise. Our primary purpose is to eliminate flux-dependent timing noise using a theoretical model proposed by Kulkarni & Romanova (2013) and to get precise parameters from XTE J1807-294. The regular pulse arrival time delay technique was first applied to obtain its initial spin parameters but the χ^2 value is as high as 1055.55 (d.o.f. = 74) which implies that the timing noise affects largely for this source. We then applied the model proposed by Kulkarni & Romanova (2013) to reduce the flux-dependent timing noise for whole data, and the χ^2 value is still as high as 1012.79 (d.o.f. = 72). The flux decay rate shows that the data can be divided into three epochs with different levels of flux-dependent timing noise affecting on the pulse phase. We then applied the model to reduce the flux-dependent timing noise for these three epochs separately and obtained a much smaller χ^2 value of 312.08 (d.o.f. = 72), which indicates the flux-dependent timing noise has been largely reduced and the optimal spin and orbital parameters are believed closer to the true values. However, there is no significant improvement in reducing flux-dependent timing noise for whole data by a single parameter, which means that model has room for improvement. Lastly, our results show that XTE J1807-294 is spin-down, which is different the previous studies.
